CyanMonkey/Petal-50M
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: CyanMonkey/Petal-50M
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-08-04 05:33:17 +08:00
commit a351509e5c
13 changed files with 81797 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

35
.gitattributes vendored Normal file
View File

@@ -0,0 +1,35 @@
*.7z filter=lfs diff=lfs merge=lfs -text
*.arrow filter=lfs diff=lfs merge=lfs -text
*.bin filter=lfs diff=lfs merge=lfs -text
*.bz2 filter=lfs diff=lfs merge=lfs -text
*.ckpt filter=lfs diff=lfs merge=lfs -text
*.ftz filter=lfs diff=lfs merge=lfs -text
*.gz filter=lfs diff=lfs merge=lfs -text
*.h5 filter=lfs diff=lfs merge=lfs -text
*.joblib filter=lfs diff=lfs merge=lfs -text
*.lfs.* filter=lfs diff=lfs merge=lfs -text
*.mlmodel filter=lfs diff=lfs merge=lfs -text
*.model filter=lfs diff=lfs merge=lfs -text
*.msgpack filter=lfs diff=lfs merge=lfs -text
*.npy filter=lfs diff=lfs merge=lfs -text
*.npz filter=lfs diff=lfs merge=lfs -text
*.onnx filter=lfs diff=lfs merge=lfs -text
*.ot filter=lfs diff=lfs merge=lfs -text
*.parquet filter=lfs diff=lfs merge=lfs -text
*.pb filter=lfs diff=lfs merge=lfs -text
*.pickle filter=lfs diff=lfs merge=lfs -text
*.pkl filter=lfs diff=lfs merge=lfs -text
*.pt filter=lfs diff=lfs merge=lfs -text
*.pth filter=lfs diff=lfs merge=lfs -text
*.rar filter=lfs diff=lfs merge=lfs -text
*.safetensors filter=lfs diff=lfs merge=lfs -text
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
*.tar.* filter=lfs diff=lfs merge=lfs -text
*.tar filter=lfs diff=lfs merge=lfs -text
*.tflite filter=lfs diff=lfs merge=lfs -text
*.tgz filter=lfs diff=lfs merge=lfs -text
*.wasm filter=lfs diff=lfs merge=lfs -text
*.xz filter=lfs diff=lfs merge=lfs -text
*.zip filter=lfs diff=lfs merge=lfs -text
*.zst filter=lfs diff=lfs merge=lfs -text
*tfevents* filter=lfs diff=lfs merge=lfs -text

40
README.md Normal file
View File

@@ -0,0 +1,40 @@
---
language: en
tags:
- causal-lm
- gqa
- rope
- swiglu
license: apache-2.0
---
# Petal-50M
Small language model (49.7M parameters) trained from scratch.
## Architecture
| Property | Value |
|---|---|
| Layers | 14 |
| Hidden size | 512 |
| Intermediate size | 1408 |
| Attention heads | 8 (GQA kv=4) |
| Max sequence length | 1024 |
| Vocab size | 16384 |
| Tied embeddings | True |
| Total parameters | 49.691M |
## Training
- Tokens seen: 12,206,075,904
- Val loss: 1.8920
- Val PPL: 6.63
## Usage
```python
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("CyanMonkey/Petal-50M")
model = AutoModelForCausalLM.from_pretrained("CyanMonkey/Petal-50M")
inputs = tokenizer("Hello", return_tensors="pt")
output = model.generate(**inputs, max_new_tokens=50)
print(tokenizer.decode(output[0], skip_special_tokens=True))
```

30
config.json Normal file
View File

@@ -0,0 +1,30 @@
{
"architectures": [
"LlamaForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"bos_token_id": 2,
"eos_token_id": 3,
"head_dim": 64,
"hidden_act": "silu",
"hidden_size": 512,
"initializer_range": 0.02,
"intermediate_size": 1408,
"max_position_embeddings": 1024,
"mlp_bias": false,
"model_type": "llama",
"num_attention_heads": 8,
"num_hidden_layers": 14,
"num_key_value_heads": 4,
"pad_token_id": 0,
"pretraining_tp": 1,
"rms_norm_eps": 1e-05,
"rope_scaling": null,
"rope_theta": 100000.0,
"tie_word_embeddings": true,
"torch_dtype": "float32",
"transformers_version": "4.40.0",
"use_cache": true,
"vocab_size": 16384
}

164
eval_results.json Normal file
View File

@@ -0,0 +1,164 @@
[
{
"task": "hellaswag",
"benchmark": "HellaSwag",
"metric": "acc_norm",
"score": 0.27972515435172274,
"shots": 0,
"runtime_sec": 204.64,
"status": "success"
},
{
"task": "piqa",
"benchmark": "PIQA",
"metric": "acc_norm",
"score": 0.5816104461371056,
"shots": 0,
"runtime_sec": 36.28,
"status": "success"
},
{
"task": "winogrande",
"benchmark": "WinoGrande",
"metric": "acc",
"score": 0.5130228887134964,
"shots": 0,
"runtime_sec": 26.49,
"status": "success"
},
{
"task": "boolq",
"benchmark": "BoolQ",
"metric": "acc",
"score": 0.6207951070336392,
"shots": 0,
"runtime_sec": 70.07,
"status": "success"
},
{
"task": "arc_easy",
"benchmark": "ARC-Easy",
"metric": "acc_norm",
"score": 0.35942760942760943,
"shots": 0,
"runtime_sec": 44.26,
"status": "success"
},
{
"task": "arc_challenge",
"benchmark": "ARC-Challenge",
"metric": "acc_norm",
"score": 0.24146757679180889,
"shots": 0,
"runtime_sec": 34.79,
"status": "success"
},
{
"task": "openbookqa",
"benchmark": "OpenBookQA",
"metric": "acc_norm",
"score": 0.274,
"shots": 0,
"runtime_sec": 26.94,
"status": "success"
},
{
"task": "commonsense_qa",
"benchmark": "CommonsenseQA",
"metric": "acc",
"score": 0.1941031941031941,
"shots": 0,
"runtime_sec": 29.94,
"status": "success"
},
{
"task": "lambada_openai",
"benchmark": "LAMBADA",
"metric": "acc",
"score": 0.19076266252668347,
"shots": 0,
"runtime_sec": 75.64,
"status": "success"
},
{
"task": "blimp",
"benchmark": "BLiMP",
"metric": "acc",
"score": 0.764268656716418,
"shots": 0,
"runtime_sec": 463.22,
"status": "success"
},
{
"task": "mmlu",
"benchmark": "MMLU",
"metric": "acc",
"score": 0.23280159521435692,
"shots": 0,
"runtime_sec": 546.26,
"status": "success"
},
{
"task": "wikitext",
"benchmark": "WikiText-2",
"metric": "word_perplexity",
"score": 55.30558314095939,
"shots": 0,
"runtime_sec": 59.4,
"status": "success"
},
{
"task": "wikitext",
"benchmark": "WikiText-2",
"metric": "byte_perplexity",
"score": 2.117905168137186,
"shots": 0,
"runtime_sec": 53.65,
"status": "success"
},
{
"task": "sciq",
"benchmark": "SciQ",
"metric": "acc_norm",
"score": 0.491,
"shots": 0,
"runtime_sec": 59.38,
"status": "success"
},
{
"task": "copa",
"benchmark": "COPA",
"metric": "acc",
"score": 0.59,
"shots": 0,
"runtime_sec": 21.9,
"status": "success"
},
{
"task": "race",
"benchmark": "RACE",
"metric": "acc",
"score": 0.25933014354066986,
"shots": 0,
"runtime_sec": 179.85,
"status": "success"
},
{
"task": "swag",
"benchmark": "SWAG",
"metric": "acc_norm",
"score": 0.4239728081575527,
"shots": 0,
"runtime_sec": 169.58,
"status": "success"
},
{
"task": "truthfulqa_mc2",
"benchmark": "TruthfulQA MC2",
"metric": "acc",
"score": 0.4401996862756814,
"shots": 0,
"runtime_sec": 74.59,
"status": "success"
}
]

4
generation_config.json Normal file
View File

@@ -0,0 +1,4 @@
{
"bos_token_id": 2,
"eos_token_id": 3
}

3
model.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:3c5f2e283c137ec35c23d5158b427c404014e08999b2e784f5aaf05e0d8238c7
size 198778664

3
resume/ckpt.pt Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:08be3988a3a4a5486fa52024451da8c22a6f0e0eeae0f9372a81c1360ff0d8bf
size 600137155

1
resume/latest_step.txt Normal file
View File

@@ -0,0 +1 @@
11652

81056
tokenizer.json Normal file
View File

File diff suppressed because it is too large Load Diff

27
tokenizer_config.json Normal file
View File

@@ -0,0 +1,27 @@
{
"backend": "tokenizers",
"bos_token": "<|bos|>",
"clean_up_tokenization_spaces": false,
"eos_token": "<|eos|>",
"extra_special_tokens": [
"<|unk_text|>",
"<|sot|>",
"<|eot|>",
"<|system|>",
"<|user|>",
"<|assistant|>",
"<|math|>",
"<|expr|>",
"<|answer|>",
"<|code|>",
"<|think|>",
"<|end_of_think|>"
],
"is_local": false,
"local_files_only": false,
"mask_token": "<|mask|>",
"model_max_length": 16384,
"pad_token": "<|pad|>",
"tokenizer_class": "TokenizersBackend",
"unk_token": "<|unk|>"
}

150
train_log.jsonl Normal file
View File

@@ -0,0 +1,150 @@
{"event": "run_start", "step": 8800, "max_steps": 11652, "model_params_M": 49.691, "eff_batch": 1024, "amp_dtype": "torch.bfloat16", "attn_backend": "chunked", "started_at": "2026-07-24T18:29:26.130247"}
{"step": 8820, "epoch": 0.757, "loss": 1.83717, "ppl": 6.279, "lr": 7.0758e-05, "lr_sched": "cosine", "grad_norm": 0.2893, "tokens": 9239408640, "tok_s": 267842.9, "elapsed_s": 131.4, "vram_gb": 2.78, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 8840, "epoch": 0.759, "loss": 1.88987, "ppl": 6.619, "lr": 7.0215e-05, "lr_sched": "cosine", "grad_norm": 0.2819, "tokens": 9260359680, "tok_s": 269322.3, "elapsed_s": 209.1, "vram_gb": 2.78, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 8860, "epoch": 0.76, "loss": 1.92752, "ppl": 6.872, "lr": 6.9675e-05, "lr_sched": "cosine", "grad_norm": 0.2926, "tokens": 9281310720, "tok_s": 269209.2, "elapsed_s": 286.8, "vram_gb": 2.78, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 8880, "epoch": 0.762, "loss": 1.82917, "ppl": 6.229, "lr": 6.9138e-05, "lr_sched": "cosine", "grad_norm": 0.3123, "tokens": 9302261760, "tok_s": 269646.1, "elapsed_s": 364.5, "vram_gb": 2.78, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 8900, "epoch": 0.764, "loss": 1.86472, "ppl": 6.454, "lr": 6.8604e-05, "lr_sched": "cosine", "grad_norm": 0.2714, "tokens": 9323212800, "tok_s": 268452.9, "elapsed_s": 442.3, "vram_gb": 2.78, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 8920, "epoch": 0.766, "loss": 1.90822, "ppl": 6.741, "lr": 6.8073e-05, "lr_sched": "cosine", "grad_norm": 0.2723, "tokens": 9344163840, "tok_s": 268963.8, "elapsed_s": 520.1, "vram_gb": 2.78, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 8940, "epoch": 0.767, "loss": 1.91169, "ppl": 6.765, "lr": 6.7545e-05, "lr_sched": "cosine", "grad_norm": 0.2922, "tokens": 9365114880, "tok_s": 268677.9, "elapsed_s": 598.0, "vram_gb": 2.78, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 8960, "epoch": 0.769, "loss": 1.80572, "ppl": 6.084, "lr": 6.702e-05, "lr_sched": "cosine", "grad_norm": 0.2728, "tokens": 9386065920, "tok_s": 268837.0, "elapsed_s": 675.6, "vram_gb": 2.78, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 8980, "epoch": 0.771, "loss": 1.88449, "ppl": 6.583, "lr": 6.6499e-05, "lr_sched": "cosine", "grad_norm": 0.2686, "tokens": 9407016960, "tok_s": 270643.9, "elapsed_s": 753.3, "vram_gb": 2.78, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9000, "epoch": 0.772, "loss": 1.85663, "ppl": 6.402, "lr": 6.598e-05, "lr_sched": "cosine", "grad_norm": 0.3064, "tokens": 9427968000, "tok_s": 269440.4, "elapsed_s": 831.0, "vram_gb": 2.78, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9020, "epoch": 0.774, "loss": 1.86381, "ppl": 6.448, "lr": 6.5465e-05, "lr_sched": "cosine", "grad_norm": 0.2849, "tokens": 9448919040, "tok_s": 269911.5, "elapsed_s": 908.8, "vram_gb": 2.78, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9040, "epoch": 0.776, "loss": 1.84236, "ppl": 6.311, "lr": 6.4952e-05, "lr_sched": "cosine", "grad_norm": 0.2711, "tokens": 9469870080, "tok_s": 269519.0, "elapsed_s": 986.7, "vram_gb": 2.78, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9060, "epoch": 0.778, "loss": 1.90693, "ppl": 6.732, "lr": 6.4443e-05, "lr_sched": "cosine", "grad_norm": 0.3108, "tokens": 9490821120, "tok_s": 268620.4, "elapsed_s": 1064.6, "vram_gb": 2.78, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9080, "epoch": 0.779, "loss": 1.84251, "ppl": 6.312, "lr": 6.3937e-05, "lr_sched": "cosine", "grad_norm": 0.3007, "tokens": 9511772160, "tok_s": 268382.2, "elapsed_s": 1142.5, "vram_gb": 2.78, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9100, "epoch": 0.781, "loss": 1.86656, "ppl": 6.466, "lr": 6.3435e-05, "lr_sched": "cosine", "grad_norm": 0.2822, "tokens": 9532723200, "tok_s": 269944.3, "elapsed_s": 1220.5, "vram_gb": 4.92, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9120, "epoch": 0.783, "loss": 1.83974, "ppl": 6.295, "lr": 6.2935e-05, "lr_sched": "cosine", "grad_norm": 0.2854, "tokens": 9553674240, "tok_s": 269876.0, "elapsed_s": 1298.2, "vram_gb": 4.92, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9140, "epoch": 0.784, "loss": 1.83865, "ppl": 6.288, "lr": 6.2439e-05, "lr_sched": "cosine", "grad_norm": 0.2851, "tokens": 9574625280, "tok_s": 268713.0, "elapsed_s": 1376.0, "vram_gb": 4.92, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9160, "epoch": 0.786, "loss": 1.86074, "ppl": 6.428, "lr": 6.1946e-05, "lr_sched": "cosine", "grad_norm": 0.2714, "tokens": 9595576320, "tok_s": 269246.5, "elapsed_s": 1453.9, "vram_gb": 4.92, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9180, "epoch": 0.788, "loss": 1.82733, "ppl": 6.217, "lr": 6.1456e-05, "lr_sched": "cosine", "grad_norm": 0.2863, "tokens": 9616527360, "tok_s": 267633.3, "elapsed_s": 1531.7, "vram_gb": 4.92, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9200, "epoch": 0.79, "loss": 1.80697, "ppl": 6.092, "lr": 6.097e-05, "lr_sched": "cosine", "grad_norm": 0.2805, "tokens": 9637478400, "tok_s": 269846.8, "elapsed_s": 1609.5, "vram_gb": 4.92, "ram_pct": 4.0, "disk_free_gb": 9223371971.07, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9200, "epoch": 0.79, "val_loss": 1.91013, "val_ppl": 6.754, "best_val_ppl": 6.754, "is_best": true}
{"step": 9220, "epoch": 0.791, "loss": 1.84067, "ppl": 6.301, "lr": 6.0487e-05, "lr_sched": "cosine", "grad_norm": 0.2881, "tokens": 9658429440, "tok_s": 269275.0, "elapsed_s": 2040.7, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9240, "epoch": 0.793, "loss": 1.80503, "ppl": 6.08, "lr": 6.0007e-05, "lr_sched": "cosine", "grad_norm": 0.2881, "tokens": 9679380480, "tok_s": 269726.3, "elapsed_s": 2118.5, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9260, "epoch": 0.795, "loss": 1.85043, "ppl": 6.363, "lr": 5.9531e-05, "lr_sched": "cosine", "grad_norm": 0.3165, "tokens": 9700331520, "tok_s": 269143.4, "elapsed_s": 2196.2, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9280, "epoch": 0.796, "loss": 1.76735, "ppl": 5.855, "lr": 5.9058e-05, "lr_sched": "cosine", "grad_norm": 0.2952, "tokens": 9721282560, "tok_s": 270518.8, "elapsed_s": 2274.0, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9300, "epoch": 0.798, "loss": 1.84474, "ppl": 6.326, "lr": 5.8588e-05, "lr_sched": "cosine", "grad_norm": 0.2775, "tokens": 9742233600, "tok_s": 268534.3, "elapsed_s": 2351.8, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9320, "epoch": 0.8, "loss": 1.80603, "ppl": 6.086, "lr": 5.8121e-05, "lr_sched": "cosine", "grad_norm": 0.2781, "tokens": 9763184640, "tok_s": 269600.7, "elapsed_s": 2429.6, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9340, "epoch": 0.802, "loss": 1.88347, "ppl": 6.576, "lr": 5.7659e-05, "lr_sched": "cosine", "grad_norm": 0.2817, "tokens": 9784135680, "tok_s": 270046.1, "elapsed_s": 2507.3, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9360, "epoch": 0.803, "loss": 1.82086, "ppl": 6.177, "lr": 5.7199e-05, "lr_sched": "cosine", "grad_norm": 0.2807, "tokens": 9805086720, "tok_s": 268721.7, "elapsed_s": 2585.2, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9380, "epoch": 0.805, "loss": 1.82677, "ppl": 6.214, "lr": 5.6743e-05, "lr_sched": "cosine", "grad_norm": 0.2882, "tokens": 9826037760, "tok_s": 269369.6, "elapsed_s": 2662.9, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9400, "epoch": 0.807, "loss": 1.85959, "ppl": 6.421, "lr": 5.629e-05, "lr_sched": "cosine", "grad_norm": 0.333, "tokens": 9846988800, "tok_s": 270002.4, "elapsed_s": 2740.7, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9420, "epoch": 0.808, "loss": 1.785, "ppl": 5.96, "lr": 5.5841e-05, "lr_sched": "cosine", "grad_norm": 0.2814, "tokens": 9867939840, "tok_s": 268378.1, "elapsed_s": 2818.7, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9440, "epoch": 0.81, "loss": 1.84568, "ppl": 6.332, "lr": 5.5395e-05, "lr_sched": "cosine", "grad_norm": 0.2887, "tokens": 9888890880, "tok_s": 270474.4, "elapsed_s": 2896.5, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9460, "epoch": 0.812, "loss": 1.86376, "ppl": 6.448, "lr": 5.4953e-05, "lr_sched": "cosine", "grad_norm": 0.3094, "tokens": 9909841920, "tok_s": 269359.7, "elapsed_s": 2974.3, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9480, "epoch": 0.814, "loss": 1.8521, "ppl": 6.373, "lr": 5.4514e-05, "lr_sched": "cosine", "grad_norm": 0.2873, "tokens": 9930792960, "tok_s": 269679.7, "elapsed_s": 3052.1, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9500, "epoch": 0.815, "loss": 1.76625, "ppl": 5.849, "lr": 5.4079e-05, "lr_sched": "cosine", "grad_norm": 0.3056, "tokens": 9951744000, "tok_s": 268422.4, "elapsed_s": 3129.9, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9520, "epoch": 0.817, "loss": 1.89656, "ppl": 6.663, "lr": 5.3647e-05, "lr_sched": "cosine", "grad_norm": 0.3015, "tokens": 9972695040, "tok_s": 269291.3, "elapsed_s": 3207.7, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9540, "epoch": 0.819, "loss": 1.80714, "ppl": 6.093, "lr": 5.3219e-05, "lr_sched": "cosine", "grad_norm": 0.2886, "tokens": 9993646080, "tok_s": 269399.2, "elapsed_s": 3285.5, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9560, "epoch": 0.82, "loss": 1.83155, "ppl": 6.244, "lr": 5.2794e-05, "lr_sched": "cosine", "grad_norm": 0.2981, "tokens": 10014597120, "tok_s": 269282.1, "elapsed_s": 3363.4, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9580, "epoch": 0.822, "loss": 1.87333, "ppl": 6.51, "lr": 5.2373e-05, "lr_sched": "cosine", "grad_norm": 0.3048, "tokens": 10035548160, "tok_s": 269561.1, "elapsed_s": 3441.2, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9600, "epoch": 0.824, "loss": 1.83243, "ppl": 6.249, "lr": 5.1955e-05, "lr_sched": "cosine", "grad_norm": 0.2648, "tokens": 10056499200, "tok_s": 269895.9, "elapsed_s": 3519.0, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371969.45, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9600, "epoch": 0.824, "val_loss": 1.905, "val_ppl": 6.719, "best_val_ppl": 6.719, "is_best": true}
{"step": 9620, "epoch": 0.826, "loss": 1.85277, "ppl": 6.377, "lr": 5.1541e-05, "lr_sched": "cosine", "grad_norm": 0.3153, "tokens": 10077450240, "tok_s": 270132.4, "elapsed_s": 3915.2, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9640, "epoch": 0.827, "loss": 1.86648, "ppl": 6.466, "lr": 5.1131e-05, "lr_sched": "cosine", "grad_norm": 0.2929, "tokens": 10098401280, "tok_s": 269661.8, "elapsed_s": 3993.0, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9660, "epoch": 0.829, "loss": 1.85813, "ppl": 6.412, "lr": 5.0724e-05, "lr_sched": "cosine", "grad_norm": 0.2854, "tokens": 10119352320, "tok_s": 268408.3, "elapsed_s": 4070.8, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9680, "epoch": 0.831, "loss": 1.83478, "ppl": 6.264, "lr": 5.0321e-05, "lr_sched": "cosine", "grad_norm": 0.2679, "tokens": 10140303360, "tok_s": 270136.8, "elapsed_s": 4148.6, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9700, "epoch": 0.832, "loss": 1.91241, "ppl": 6.769, "lr": 4.9921e-05, "lr_sched": "cosine", "grad_norm": 0.3052, "tokens": 10161254400, "tok_s": 268227.7, "elapsed_s": 4226.4, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9720, "epoch": 0.834, "loss": 1.88976, "ppl": 6.618, "lr": 4.9526e-05, "lr_sched": "cosine", "grad_norm": 0.2979, "tokens": 10182205440, "tok_s": 270479.4, "elapsed_s": 4304.1, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9740, "epoch": 0.836, "loss": 1.85868, "ppl": 6.415, "lr": 4.9133e-05, "lr_sched": "cosine", "grad_norm": 0.2899, "tokens": 10203156480, "tok_s": 269697.5, "elapsed_s": 4381.9, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9760, "epoch": 0.838, "loss": 1.88569, "ppl": 6.591, "lr": 4.8745e-05, "lr_sched": "cosine", "grad_norm": 0.2767, "tokens": 10224107520, "tok_s": 269154.8, "elapsed_s": 4459.7, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9780, "epoch": 0.839, "loss": 1.83973, "ppl": 6.295, "lr": 4.836e-05, "lr_sched": "cosine", "grad_norm": 0.3132, "tokens": 10245058560, "tok_s": 268768.6, "elapsed_s": 4537.5, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9800, "epoch": 0.841, "loss": 1.84633, "ppl": 6.337, "lr": 4.7979e-05, "lr_sched": "cosine", "grad_norm": 0.2694, "tokens": 10266009600, "tok_s": 269045.9, "elapsed_s": 4615.3, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9820, "epoch": 0.843, "loss": 1.85392, "ppl": 6.385, "lr": 4.7602e-05, "lr_sched": "cosine", "grad_norm": 0.2859, "tokens": 10286960640, "tok_s": 269634.0, "elapsed_s": 4693.1, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9840, "epoch": 0.844, "loss": 1.83725, "ppl": 6.279, "lr": 4.7228e-05, "lr_sched": "cosine", "grad_norm": 0.2653, "tokens": 10307911680, "tok_s": 269518.0, "elapsed_s": 4770.8, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9860, "epoch": 0.846, "loss": 1.81998, "ppl": 6.172, "lr": 4.6858e-05, "lr_sched": "cosine", "grad_norm": 0.3287, "tokens": 10328862720, "tok_s": 268917.3, "elapsed_s": 4848.5, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9880, "epoch": 0.848, "loss": 1.83675, "ppl": 6.276, "lr": 4.6492e-05, "lr_sched": "cosine", "grad_norm": 0.2987, "tokens": 10349813760, "tok_s": 268454.0, "elapsed_s": 4926.3, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9900, "epoch": 0.85, "loss": 1.84696, "ppl": 6.34, "lr": 4.6129e-05, "lr_sched": "cosine", "grad_norm": 0.2693, "tokens": 10370764800, "tok_s": 268964.2, "elapsed_s": 5004.1, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9920, "epoch": 0.851, "loss": 1.86191, "ppl": 6.436, "lr": 4.5771e-05, "lr_sched": "cosine", "grad_norm": 0.2919, "tokens": 10391715840, "tok_s": 270257.6, "elapsed_s": 5081.8, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9940, "epoch": 0.853, "loss": 1.85315, "ppl": 6.38, "lr": 4.5416e-05, "lr_sched": "cosine", "grad_norm": 0.3028, "tokens": 10412666880, "tok_s": 270312.6, "elapsed_s": 5159.5, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9960, "epoch": 0.855, "loss": 1.8162, "ppl": 6.148, "lr": 4.5065e-05, "lr_sched": "cosine", "grad_norm": 0.2705, "tokens": 10433617920, "tok_s": 268821.7, "elapsed_s": 5237.9, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 9980, "epoch": 0.857, "loss": 1.85228, "ppl": 6.374, "lr": 4.4718e-05, "lr_sched": "cosine", "grad_norm": 0.302, "tokens": 10454568960, "tok_s": 268223.6, "elapsed_s": 5316.1, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10000, "epoch": 0.858, "loss": 1.90574, "ppl": 6.724, "lr": 4.4374e-05, "lr_sched": "cosine", "grad_norm": 0.2711, "tokens": 10475520000, "tok_s": 268494.0, "elapsed_s": 5394.3, "vram_gb": 2.78, "ram_pct": 4.1, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10000, "epoch": 0.858, "val_loss": 1.89983, "val_ppl": 6.685, "best_val_ppl": 6.685, "is_best": true}
{"step": 10020, "epoch": 0.86, "loss": 1.84914, "ppl": 6.354, "lr": 4.4035e-05, "lr_sched": "cosine", "grad_norm": 0.2677, "tokens": 10496471040, "tok_s": 268508.1, "elapsed_s": 5791.5, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10040, "epoch": 0.862, "loss": 1.84401, "ppl": 6.322, "lr": 4.3699e-05, "lr_sched": "cosine", "grad_norm": 0.2547, "tokens": 10517422080, "tok_s": 270180.0, "elapsed_s": 5869.2, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10060, "epoch": 0.863, "loss": 1.86883, "ppl": 6.481, "lr": 4.3367e-05, "lr_sched": "cosine", "grad_norm": 0.2798, "tokens": 10538373120, "tok_s": 270413.4, "elapsed_s": 5946.9, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10080, "epoch": 0.865, "loss": 1.84381, "ppl": 6.321, "lr": 4.3039e-05, "lr_sched": "cosine", "grad_norm": 0.3107, "tokens": 10559324160, "tok_s": 268577.4, "elapsed_s": 6024.7, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10100, "epoch": 0.867, "loss": 1.99061, "ppl": 7.32, "lr": 4.2715e-05, "lr_sched": "cosine", "grad_norm": 0.3802, "tokens": 10580275200, "tok_s": 269402.6, "elapsed_s": 6102.6, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10120, "epoch": 0.869, "loss": 1.75485, "ppl": 5.783, "lr": 4.2394e-05, "lr_sched": "cosine", "grad_norm": 0.2806, "tokens": 10601226240, "tok_s": 267510.9, "elapsed_s": 6180.5, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10140, "epoch": 0.87, "loss": 1.99572, "ppl": 7.357, "lr": 4.2078e-05, "lr_sched": "cosine", "grad_norm": 0.8364, "tokens": 10622177280, "tok_s": 268305.2, "elapsed_s": 6258.7, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10160, "epoch": 0.872, "loss": 1.86967, "ppl": 6.486, "lr": 4.1765e-05, "lr_sched": "cosine", "grad_norm": 0.3092, "tokens": 10643128320, "tok_s": 268552.8, "elapsed_s": 6336.6, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10180, "epoch": 0.874, "loss": 1.83294, "ppl": 6.252, "lr": 4.1457e-05, "lr_sched": "cosine", "grad_norm": 0.2595, "tokens": 10664079360, "tok_s": 269278.4, "elapsed_s": 6414.5, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10200, "epoch": 0.875, "loss": 1.85806, "ppl": 6.411, "lr": 4.1152e-05, "lr_sched": "cosine", "grad_norm": 0.3023, "tokens": 10685030400, "tok_s": 267794.3, "elapsed_s": 6492.5, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10220, "epoch": 0.877, "loss": 1.80467, "ppl": 6.078, "lr": 4.0851e-05, "lr_sched": "cosine", "grad_norm": 0.293, "tokens": 10705981440, "tok_s": 266895.5, "elapsed_s": 6570.5, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10240, "epoch": 0.879, "loss": 1.78058, "ppl": 5.933, "lr": 4.0555e-05, "lr_sched": "cosine", "grad_norm": 0.2668, "tokens": 10726932480, "tok_s": 266388.7, "elapsed_s": 6648.8, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10260, "epoch": 0.881, "loss": 1.79627, "ppl": 6.027, "lr": 4.0262e-05, "lr_sched": "cosine", "grad_norm": 0.2867, "tokens": 10747883520, "tok_s": 267577.8, "elapsed_s": 6727.3, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10280, "epoch": 0.882, "loss": 1.8656, "ppl": 6.46, "lr": 3.9973e-05, "lr_sched": "cosine", "grad_norm": 0.3036, "tokens": 10768834560, "tok_s": 267173.3, "elapsed_s": 6805.6, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10300, "epoch": 0.884, "loss": 1.78567, "ppl": 5.964, "lr": 3.9688e-05, "lr_sched": "cosine", "grad_norm": 0.29, "tokens": 10789785600, "tok_s": 269065.3, "elapsed_s": 6883.9, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10320, "epoch": 0.886, "loss": 1.80172, "ppl": 6.06, "lr": 3.9407e-05, "lr_sched": "cosine", "grad_norm": 0.2759, "tokens": 10810736640, "tok_s": 268775.9, "elapsed_s": 6961.9, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10340, "epoch": 0.887, "loss": 1.89481, "ppl": 6.651, "lr": 3.913e-05, "lr_sched": "cosine", "grad_norm": 0.2805, "tokens": 10831687680, "tok_s": 269707.8, "elapsed_s": 7039.9, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10360, "epoch": 0.889, "loss": 1.86798, "ppl": 6.475, "lr": 3.8857e-05, "lr_sched": "cosine", "grad_norm": 0.2885, "tokens": 10852638720, "tok_s": 266906.9, "elapsed_s": 7117.9, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10380, "epoch": 0.891, "loss": 1.80931, "ppl": 6.106, "lr": 3.8588e-05, "lr_sched": "cosine", "grad_norm": 0.2764, "tokens": 10873589760, "tok_s": 268031.8, "elapsed_s": 7195.8, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10400, "epoch": 0.893, "loss": 1.86338, "ppl": 6.446, "lr": 3.8323e-05, "lr_sched": "cosine", "grad_norm": 0.2763, "tokens": 10894540800, "tok_s": 269515.4, "elapsed_s": 7273.6, "vram_gb": 2.78, "ram_pct": 4.2, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10400, "epoch": 0.893, "val_loss": 1.896, "val_ppl": 6.659, "best_val_ppl": 6.659, "is_best": true}
{"step": 10420, "epoch": 0.894, "loss": 1.83668, "ppl": 6.276, "lr": 3.8062e-05, "lr_sched": "cosine", "grad_norm": 0.2941, "tokens": 10915491840, "tok_s": 269255.4, "elapsed_s": 7673.1, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10440, "epoch": 0.896, "loss": 1.87926, "ppl": 6.549, "lr": 3.7805e-05, "lr_sched": "cosine", "grad_norm": 0.2699, "tokens": 10936442880, "tok_s": 268953.8, "elapsed_s": 7751.2, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10460, "epoch": 0.898, "loss": 1.83896, "ppl": 6.29, "lr": 3.7552e-05, "lr_sched": "cosine", "grad_norm": 0.2742, "tokens": 10957393920, "tok_s": 267855.8, "elapsed_s": 7829.3, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10480, "epoch": 0.899, "loss": 1.87086, "ppl": 6.494, "lr": 3.7304e-05, "lr_sched": "cosine", "grad_norm": 0.2886, "tokens": 10978344960, "tok_s": 268423.4, "elapsed_s": 7907.5, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10500, "epoch": 0.901, "loss": 1.8841, "ppl": 6.58, "lr": 3.7059e-05, "lr_sched": "cosine", "grad_norm": 0.2775, "tokens": 10999296000, "tok_s": 266778.4, "elapsed_s": 7985.6, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10520, "epoch": 0.903, "loss": 1.9385, "ppl": 6.948, "lr": 3.6818e-05, "lr_sched": "cosine", "grad_norm": 0.4248, "tokens": 11020247040, "tok_s": 268818.6, "elapsed_s": 8063.7, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10540, "epoch": 0.905, "loss": 1.82055, "ppl": 6.175, "lr": 3.6582e-05, "lr_sched": "cosine", "grad_norm": 0.3095, "tokens": 11041198080, "tok_s": 266669.7, "elapsed_s": 8141.6, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10560, "epoch": 0.906, "loss": 1.87304, "ppl": 6.508, "lr": 3.6349e-05, "lr_sched": "cosine", "grad_norm": 0.3212, "tokens": 11062149120, "tok_s": 268009.9, "elapsed_s": 8219.6, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10580, "epoch": 0.908, "loss": 1.8236, "ppl": 6.194, "lr": 3.6121e-05, "lr_sched": "cosine", "grad_norm": 0.3018, "tokens": 11083100160, "tok_s": 269219.3, "elapsed_s": 8297.6, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10600, "epoch": 0.91, "loss": 1.74339, "ppl": 5.717, "lr": 3.5896e-05, "lr_sched": "cosine", "grad_norm": 0.3303, "tokens": 11104051200, "tok_s": 269249.3, "elapsed_s": 8375.5, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10620, "epoch": 0.911, "loss": 1.84348, "ppl": 6.318, "lr": 3.5676e-05, "lr_sched": "cosine", "grad_norm": 0.3023, "tokens": 11125002240, "tok_s": 268201.1, "elapsed_s": 8453.7, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10640, "epoch": 0.913, "loss": 1.83174, "ppl": 6.245, "lr": 3.546e-05, "lr_sched": "cosine", "grad_norm": 0.3136, "tokens": 11145953280, "tok_s": 268062.0, "elapsed_s": 8531.7, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10660, "epoch": 0.915, "loss": 1.86049, "ppl": 6.427, "lr": 3.5248e-05, "lr_sched": "cosine", "grad_norm": 0.2733, "tokens": 11166904320, "tok_s": 267553.6, "elapsed_s": 8609.7, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10680, "epoch": 0.917, "loss": 1.81944, "ppl": 6.168, "lr": 3.504e-05, "lr_sched": "cosine", "grad_norm": 0.3243, "tokens": 11187855360, "tok_s": 269257.8, "elapsed_s": 8687.8, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10700, "epoch": 0.918, "loss": 1.84637, "ppl": 6.337, "lr": 3.4836e-05, "lr_sched": "cosine", "grad_norm": 0.3338, "tokens": 11208806400, "tok_s": 268388.6, "elapsed_s": 8765.9, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10720, "epoch": 0.92, "loss": 1.82114, "ppl": 6.179, "lr": 3.4636e-05, "lr_sched": "cosine", "grad_norm": 0.3199, "tokens": 11229757440, "tok_s": 268854.1, "elapsed_s": 8844.0, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10740, "epoch": 0.922, "loss": 1.78886, "ppl": 5.983, "lr": 3.4441e-05, "lr_sched": "cosine", "grad_norm": 0.3535, "tokens": 11250708480, "tok_s": 269901.7, "elapsed_s": 8921.9, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10760, "epoch": 0.923, "loss": 1.86321, "ppl": 6.444, "lr": 3.4249e-05, "lr_sched": "cosine", "grad_norm": 0.2679, "tokens": 11271659520, "tok_s": 269196.4, "elapsed_s": 8999.9, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10780, "epoch": 0.925, "loss": 1.92557, "ppl": 6.859, "lr": 3.4062e-05, "lr_sched": "cosine", "grad_norm": 0.2609, "tokens": 11292610560, "tok_s": 268816.8, "elapsed_s": 9077.9, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10800, "epoch": 0.927, "loss": 1.88716, "ppl": 6.601, "lr": 3.3879e-05, "lr_sched": "cosine", "grad_norm": 0.2489, "tokens": 11313561600, "tok_s": 268131.0, "elapsed_s": 9155.8, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.85, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10800, "epoch": 0.927, "val_loss": 1.89445, "val_ppl": 6.649, "best_val_ppl": 6.649, "is_best": true}
{"step": 10820, "epoch": 0.929, "loss": 1.8496, "ppl": 6.357, "lr": 3.37e-05, "lr_sched": "cosine", "grad_norm": 0.2849, "tokens": 11334512640, "tok_s": 269642.8, "elapsed_s": 9554.7, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10840, "epoch": 0.93, "loss": 1.90154, "ppl": 6.696, "lr": 3.3525e-05, "lr_sched": "cosine", "grad_norm": 0.2654, "tokens": 11355463680, "tok_s": 267937.2, "elapsed_s": 9632.5, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10860, "epoch": 0.932, "loss": 1.84298, "ppl": 6.315, "lr": 3.3355e-05, "lr_sched": "cosine", "grad_norm": 0.2703, "tokens": 11376414720, "tok_s": 268045.3, "elapsed_s": 9710.3, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10880, "epoch": 0.934, "loss": 1.88474, "ppl": 6.585, "lr": 3.3188e-05, "lr_sched": "cosine", "grad_norm": 0.2822, "tokens": 11397365760, "tok_s": 268891.2, "elapsed_s": 9788.0, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10900, "epoch": 0.935, "loss": 1.86963, "ppl": 6.486, "lr": 3.3026e-05, "lr_sched": "cosine", "grad_norm": 0.2737, "tokens": 11418316800, "tok_s": 269174.0, "elapsed_s": 9865.7, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10920, "epoch": 0.937, "loss": 1.86235, "ppl": 6.439, "lr": 3.2868e-05, "lr_sched": "cosine", "grad_norm": 0.2685, "tokens": 11439267840, "tok_s": 268914.4, "elapsed_s": 9943.6, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10940, "epoch": 0.939, "loss": 1.82836, "ppl": 6.224, "lr": 3.2714e-05, "lr_sched": "cosine", "grad_norm": 0.2662, "tokens": 11460218880, "tok_s": 268764.1, "elapsed_s": 10021.5, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10960, "epoch": 0.941, "loss": 1.82516, "ppl": 6.204, "lr": 3.2564e-05, "lr_sched": "cosine", "grad_norm": 0.2566, "tokens": 11481169920, "tok_s": 269449.4, "elapsed_s": 10099.3, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 10980, "epoch": 0.942, "loss": 1.78976, "ppl": 5.988, "lr": 3.2419e-05, "lr_sched": "cosine", "grad_norm": 0.2703, "tokens": 11502120960, "tok_s": 269021.7, "elapsed_s": 10177.1, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11000, "epoch": 0.944, "loss": 1.87938, "ppl": 6.549, "lr": 3.2278e-05, "lr_sched": "cosine", "grad_norm": 0.2862, "tokens": 11523072000, "tok_s": 269527.7, "elapsed_s": 10254.9, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11020, "epoch": 0.946, "loss": 1.86884, "ppl": 6.481, "lr": 3.2141e-05, "lr_sched": "cosine", "grad_norm": 0.2893, "tokens": 11544023040, "tok_s": 269423.7, "elapsed_s": 10332.7, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11040, "epoch": 0.947, "loss": 1.77045, "ppl": 5.874, "lr": 3.2008e-05, "lr_sched": "cosine", "grad_norm": 0.2762, "tokens": 11564974080, "tok_s": 268932.0, "elapsed_s": 10410.5, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11060, "epoch": 0.949, "loss": 1.8436, "ppl": 6.319, "lr": 3.1879e-05, "lr_sched": "cosine", "grad_norm": 0.2839, "tokens": 11585925120, "tok_s": 268462.4, "elapsed_s": 10488.3, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11080, "epoch": 0.951, "loss": 1.83254, "ppl": 6.25, "lr": 3.1755e-05, "lr_sched": "cosine", "grad_norm": 0.2593, "tokens": 11606876160, "tok_s": 269721.7, "elapsed_s": 10566.1, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11100, "epoch": 0.953, "loss": 1.81423, "ppl": 6.136, "lr": 3.1635e-05, "lr_sched": "cosine", "grad_norm": 0.2913, "tokens": 11627827200, "tok_s": 268992.3, "elapsed_s": 10643.9, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11120, "epoch": 0.954, "loss": 1.83394, "ppl": 6.259, "lr": 3.1519e-05, "lr_sched": "cosine", "grad_norm": 0.2901, "tokens": 11648778240, "tok_s": 269165.8, "elapsed_s": 10721.8, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11140, "epoch": 0.956, "loss": 1.84829, "ppl": 6.349, "lr": 3.1407e-05, "lr_sched": "cosine", "grad_norm": 0.2954, "tokens": 11669729280, "tok_s": 268702.6, "elapsed_s": 10799.7, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11160, "epoch": 0.958, "loss": 1.88456, "ppl": 6.583, "lr": 3.13e-05, "lr_sched": "cosine", "grad_norm": 0.2636, "tokens": 11690680320, "tok_s": 268959.7, "elapsed_s": 10877.6, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11180, "epoch": 0.959, "loss": 1.87637, "ppl": 6.53, "lr": 3.1197e-05, "lr_sched": "cosine", "grad_norm": 0.2698, "tokens": 11711631360, "tok_s": 269320.6, "elapsed_s": 10955.4, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11200, "epoch": 0.961, "loss": 1.88903, "ppl": 6.613, "lr": 3.1098e-05, "lr_sched": "cosine", "grad_norm": 0.263, "tokens": 11732582400, "tok_s": 269877.5, "elapsed_s": 11033.2, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11200, "epoch": 0.961, "val_loss": 1.89248, "val_ppl": 6.636, "best_val_ppl": 6.636, "is_best": true}
{"step": 11220, "epoch": 0.963, "loss": 1.92929, "ppl": 6.885, "lr": 3.1003e-05, "lr_sched": "cosine", "grad_norm": 0.2958, "tokens": 11753533440, "tok_s": 268538.6, "elapsed_s": 11428.5, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11240, "epoch": 0.965, "loss": 1.89082, "ppl": 6.625, "lr": 3.0913e-05, "lr_sched": "cosine", "grad_norm": 0.2614, "tokens": 11774484480, "tok_s": 269335.0, "elapsed_s": 11506.2, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11260, "epoch": 0.966, "loss": 1.85848, "ppl": 6.414, "lr": 3.0826e-05, "lr_sched": "cosine", "grad_norm": 0.2983, "tokens": 11795435520, "tok_s": 268240.1, "elapsed_s": 11584.0, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11280, "epoch": 0.968, "loss": 1.84137, "ppl": 6.305, "lr": 3.0745e-05, "lr_sched": "cosine", "grad_norm": 0.2807, "tokens": 11816386560, "tok_s": 268500.9, "elapsed_s": 11661.9, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11300, "epoch": 0.97, "loss": 1.71302, "ppl": 5.546, "lr": 3.0667e-05, "lr_sched": "cosine", "grad_norm": 0.3021, "tokens": 11837337600, "tok_s": 268683.8, "elapsed_s": 11739.6, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11320, "epoch": 0.972, "loss": 1.84834, "ppl": 6.349, "lr": 3.0594e-05, "lr_sched": "cosine", "grad_norm": 0.3048, "tokens": 11858288640, "tok_s": 270240.1, "elapsed_s": 11817.3, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11340, "epoch": 0.973, "loss": 1.85074, "ppl": 6.365, "lr": 3.0524e-05, "lr_sched": "cosine", "grad_norm": 0.2703, "tokens": 11879239680, "tok_s": 269813.0, "elapsed_s": 11895.0, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11360, "epoch": 0.975, "loss": 1.39704, "ppl": 4.043, "lr": 3.046e-05, "lr_sched": "cosine", "grad_norm": 0.3029, "tokens": 11900190720, "tok_s": 268804.5, "elapsed_s": 11972.8, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11380, "epoch": 0.977, "loss": 1.87909, "ppl": 6.548, "lr": 3.0399e-05, "lr_sched": "cosine", "grad_norm": 0.3174, "tokens": 11921141760, "tok_s": 270384.7, "elapsed_s": 12050.4, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11400, "epoch": 0.978, "loss": 1.8049, "ppl": 6.079, "lr": 3.0343e-05, "lr_sched": "cosine", "grad_norm": 0.2804, "tokens": 11942092800, "tok_s": 270251.3, "elapsed_s": 12128.2, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11420, "epoch": 0.98, "loss": 1.78632, "ppl": 5.967, "lr": 3.0291e-05, "lr_sched": "cosine", "grad_norm": 0.2868, "tokens": 11963043840, "tok_s": 269264.0, "elapsed_s": 12205.9, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11440, "epoch": 0.982, "loss": 1.81081, "ppl": 6.115, "lr": 3.0243e-05, "lr_sched": "cosine", "grad_norm": 0.2888, "tokens": 11983994880, "tok_s": 269880.8, "elapsed_s": 12283.7, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11460, "epoch": 0.984, "loss": 1.86827, "ppl": 6.477, "lr": 3.0199e-05, "lr_sched": "cosine", "grad_norm": 0.3069, "tokens": 12004945920, "tok_s": 268846.8, "elapsed_s": 12361.4, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11480, "epoch": 0.985, "loss": 1.81303, "ppl": 6.129, "lr": 3.016e-05, "lr_sched": "cosine", "grad_norm": 0.3257, "tokens": 12025896960, "tok_s": 268923.7, "elapsed_s": 12439.2, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11500, "epoch": 0.987, "loss": 1.87356, "ppl": 6.511, "lr": 3.0125e-05, "lr_sched": "cosine", "grad_norm": 0.2923, "tokens": 12046848000, "tok_s": 269397.7, "elapsed_s": 12516.9, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11520, "epoch": 0.989, "loss": 1.81918, "ppl": 6.167, "lr": 3.0095e-05, "lr_sched": "cosine", "grad_norm": 0.2609, "tokens": 12067799040, "tok_s": 269716.9, "elapsed_s": 12594.7, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11540, "epoch": 0.99, "loss": 1.83873, "ppl": 6.289, "lr": 3.0068e-05, "lr_sched": "cosine", "grad_norm": 0.2673, "tokens": 12088750080, "tok_s": 268892.8, "elapsed_s": 12672.4, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11560, "epoch": 0.992, "loss": 1.8156, "ppl": 6.145, "lr": 3.0046e-05, "lr_sched": "cosine", "grad_norm": 0.3319, "tokens": 12109701120, "tok_s": 269900.1, "elapsed_s": 12750.1, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11580, "epoch": 0.994, "loss": 1.8363, "ppl": 6.273, "lr": 3.0029e-05, "lr_sched": "cosine", "grad_norm": 0.2887, "tokens": 12130652160, "tok_s": 269548.5, "elapsed_s": 12827.9, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11600, "epoch": 0.996, "loss": 1.82102, "ppl": 6.178, "lr": 3.0015e-05, "lr_sched": "cosine", "grad_norm": 0.2774, "tokens": 12151603200, "tok_s": 270289.1, "elapsed_s": 12905.7, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11600, "epoch": 0.996, "val_loss": 1.89176, "val_ppl": 6.631, "best_val_ppl": 6.631, "is_best": true}
{"step": 11620, "epoch": 0.997, "loss": 1.84643, "ppl": 6.337, "lr": 3.0006e-05, "lr_sched": "cosine", "grad_norm": 0.3379, "tokens": 12172554240, "tok_s": 270729.8, "elapsed_s": 13300.7, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}
{"step": 11640, "epoch": 0.999, "loss": 1.79849, "ppl": 6.041, "lr": 3.0001e-05, "lr_sched": "cosine", "grad_norm": 0.3078, "tokens": 12193505280, "tok_s": 268458.1, "elapsed_s": 13378.6, "vram_gb": 2.78, "ram_pct": 4.4, "disk_free_gb": 9223371968.84, "attn_backend": "chunked", "amp_dtype": "torch.bfloat16"}

276
train_log.log Normal file
View File

@@ -0,0 +1,276 @@
2026-07-24 18:24:24,991 | HTTP Request: GET https://huggingface.co/api/agent-harnesses "HTTP/1.1 200 OK"
2026-07-24 18:24:25,017 | HTTP Request: GET https://huggingface.co/api/whoami-v2 "HTTP/1.1 200 OK"
2026-07-24 18:24:25,069 | HTTP Request: HEAD https://huggingface.co/GODELEV/TOK-16K/resolve/main/config.json "HTTP/1.1 404 Not Found"
2026-07-24 18:24:25,099 | HTTP Request: HEAD https://huggingface.co/GODELEV/TOK-16K/resolve/main/config.json "HTTP/1.1 404 Not Found"
2026-07-24 18:24:25,130 | HTTP Request: HEAD https://huggingface.co/GODELEV/TOK-16K/resolve/main/tokenizer_config.json "HTTP/1.1 307 Temporary Redirect"
2026-07-24 18:24:25,141 | HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/GODELEV/TOK-16K/75fdb3c4114e2b79cac5e43fc8c4a8d5e48acbb3/tokenizer_config.json "HTTP/1.1 200 OK"
2026-07-24 18:24:25,153 | HTTP Request: GET https://huggingface.co/api/resolve-cache/models/GODELEV/TOK-16K/75fdb3c4114e2b79cac5e43fc8c4a8d5e48acbb3/tokenizer_config.json "HTTP/1.1 200 OK"
2026-07-24 18:24:25,184 | HTTP Request: GET https://huggingface.co/api/models/GODELEV/TOK-16K/tree/main/additional_chat_templates?recursive=false&expand=false "HTTP/1.1 404 Not Found"
2026-07-24 18:24:25,214 | HTTP Request: GET https://huggingface.co/api/models/GODELEV/TOK-16K/tree/main?recursive=true&expand=false "HTTP/1.1 200 OK"
2026-07-24 18:24:25,241 | HTTP Request: HEAD https://huggingface.co/GODELEV/TOK-16K/resolve/main/tokenizer.json "HTTP/1.1 307 Temporary Redirect"
2026-07-24 18:24:25,252 | HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/GODELEV/TOK-16K/75fdb3c4114e2b79cac5e43fc8c4a8d5e48acbb3/tokenizer.json "HTTP/1.1 200 OK"
2026-07-24 18:24:25,263 | HTTP Request: GET https://huggingface.co/api/resolve-cache/models/GODELEV/TOK-16K/75fdb3c4114e2b79cac5e43fc8c4a8d5e48acbb3/tokenizer.json "HTTP/1.1 200 OK"
2026-07-24 18:24:25,309 | HTTP Request: HEAD https://huggingface.co/GODELEV/TOK-16K/resolve/main/tokenizer.model "HTTP/1.1 404 Not Found"
2026-07-24 18:24:25,339 | HTTP Request: HEAD https://huggingface.co/GODELEV/TOK-16K/resolve/main/added_tokens.json "HTTP/1.1 404 Not Found"
2026-07-24 18:24:25,371 | HTTP Request: HEAD https://huggingface.co/GODELEV/TOK-16K/resolve/main/special_tokens_map.json "HTTP/1.1 307 Temporary Redirect"
2026-07-24 18:24:25,382 | HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/GODELEV/TOK-16K/75fdb3c4114e2b79cac5e43fc8c4a8d5e48acbb3/special_tokens_map.json "HTTP/1.1 200 OK"
2026-07-24 18:24:25,393 | HTTP Request: GET https://huggingface.co/api/resolve-cache/models/GODELEV/TOK-16K/75fdb3c4114e2b79cac5e43fc8c4a8d5e48acbb3/special_tokens_map.json "HTTP/1.1 200 OK"
2026-07-24 18:24:25,422 | HTTP Request: HEAD https://huggingface.co/GODELEV/TOK-16K/resolve/main/chat_template.jinja "HTTP/1.1 404 Not Found"
2026-07-24 18:24:25,534 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/main/README.md "HTTP/1.1 404 Not Found"
2026-07-24 18:24:25,566 | HTTP Request: GET https://huggingface.co/api/datasets/CyanMonkey/Dataset-1-16-1024 "HTTP/1.1 200 OK"
2026-07-24 18:24:25,596 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/Dataset-1-16-1024.py "HTTP/1.1 404 Not Found"
2026-07-24 18:24:25,653 | HTTP Request: HEAD https://s3.amazonaws.com/datasets.huggingface.co/datasets/datasets/CyanMonkey/Dataset-1-16-1024/CyanMonkey/Dataset-1-16-1024.py "HTTP/1.1 404 Not Found"
2026-07-24 18:24:25,681 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/README.md "HTTP/1.1 404 Not Found"
2026-07-24 18:24:25,710 | HTTP Request: GET https://huggingface.co/api/datasets/CyanMonkey/Dataset-1-16-1024/revision/0c91017612a0ade103e68623e6db22637984be88 "HTTP/1.1 200 OK"
2026-07-24 18:24:25,736 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/.huggingface.yaml "HTTP/1.1 404 Not Found"
2026-07-24 18:24:25,822 | HTTP Request: GET https://datasets-server.huggingface.co/info?dataset=CyanMonkey/Dataset-1-16-1024 "HTTP/1.1 200 OK"
2026-07-24 18:24:25,851 | HTTP Request: GET https://huggingface.co/api/datasets/CyanMonkey/Dataset-1-16-1024/tree/0c91017612a0ade103e68623e6db22637984be88/data?recursive=true&expand=false "HTTP/1.1 200 OK"
2026-07-24 18:24:25,897 | HTTP Request: GET https://huggingface.co/api/datasets/CyanMonkey/Dataset-1-16-1024/tree/0c91017612a0ade103e68623e6db22637984be88?recursive=false&expand=false "HTTP/1.1 200 OK"
2026-07-24 18:24:25,931 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/dataset_infos.json "HTTP/1.1 404 Not Found"
2026-07-24 18:24:25,986 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00000-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:24:32,406 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00001-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:24:37,972 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00002-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:24:44,092 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00003-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:24:49,763 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00004-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:24:55,377 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00005-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:25:00,890 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00006-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:25:07,243 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00007-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:25:13,626 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00008-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:25:19,362 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00009-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:25:25,777 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00010-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:25:30,212 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00011-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:25:36,531 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00012-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:25:41,152 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00013-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:25:47,711 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00014-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:25:53,464 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00015-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:25:58,166 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00016-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:26:04,941 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00017-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:26:10,935 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00018-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:26:15,083 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00019-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:26:20,897 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00020-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:26:25,774 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00021-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:26:32,577 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00022-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:26:39,394 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00023-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:26:45,748 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00024-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:26:51,669 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00025-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:26:57,901 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00026-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:27:05,186 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00027-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:27:11,699 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00028-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:27:19,442 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/train-00029-of-00030.arrow "HTTP/1.1 302 Found"
2026-07-24 18:27:26,293 | HTTP Request: HEAD https://huggingface.co/datasets/CyanMonkey/Dataset-1-16-1024/resolve/0c91017612a0ade103e68623e6db22637984be88/data/val-00000-of-00001.arrow "HTTP/1.1 302 Found"
2026-07-24 18:29:17,865 | Data loaded train=11,931,626 val=200,000 steps_per_epoch=11,652 total_epochs=1.00
2026-07-24 18:29:19,261 | Model 49.691M params | float32 | amp=torch.bfloat16 | compiled=True | attn=chunked
2026-07-24 18:29:19,338 | HTTP Request: HEAD https://huggingface.co/CyanMonkey/Petal-50M/resolve/main/resume/latest_step.txt "HTTP/1.1 307 Temporary Redirect"
2026-07-24 18:29:19,350 | HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/CyanMonkey/Petal-50M/320b1e3e83ff8663c444de50aff846230650e989/resume%2Flatest_step.txt "HTTP/1.1 200 OK"
2026-07-24 18:29:19,363 | HTTP Request: GET https://huggingface.co/api/resolve-cache/models/CyanMonkey/Petal-50M/320b1e3e83ff8663c444de50aff846230650e989/resume%2Flatest_step.txt "HTTP/1.1 200 OK"
2026-07-24 18:29:19,408 | HTTP Request: HEAD https://huggingface.co/CyanMonkey/Petal-50M/resolve/main/resume/ckpt.pt "HTTP/1.1 302 Found"
2026-07-24 18:29:26,123 | Resumed step=8800 tokens=9218457600 samples=9011200
2026-07-24 18:31:37,533 | step= 8820 | epoch=0.76 | loss=1.8372 | ppl=6.28 | lr=7.08e-05 | grad=0.289 | tok/s=267,843 | ETA=3:04:36 | VRAM=2.8GB | RAM=4%
2026-07-24 18:32:55,278 | step= 8840 | epoch=0.76 | loss=1.8899 | ppl=6.62 | lr=7.02e-05 | grad=0.282 | tok/s=269,322 | ETA=3:02:17 | VRAM=2.8GB | RAM=4%
2026-07-24 18:34:12,977 | step= 8860 | epoch=0.76 | loss=1.9275 | ppl=6.87 | lr=6.97e-05 | grad=0.293 | tok/s=269,209 | ETA=3:01:04 | VRAM=2.8GB | RAM=4%
2026-07-24 18:35:30,680 | step= 8880 | epoch=0.76 | loss=1.8292 | ppl=6.23 | lr=6.91e-05 | grad=0.312 | tok/s=269,646 | ETA=2:59:28 | VRAM=2.8GB | RAM=4%
2026-07-24 18:36:48,478 | step= 8900 | epoch=0.76 | loss=1.8647 | ppl=6.45 | lr=6.86e-05 | grad=0.271 | tok/s=268,453 | ETA=2:58:58 | VRAM=2.8GB | RAM=4%
2026-07-24 18:38:06,220 | step= 8920 | epoch=0.77 | loss=1.9082 | ppl=6.74 | lr=6.81e-05 | grad=0.272 | tok/s=268,964 | ETA=2:57:20 | VRAM=2.8GB | RAM=4%
2026-07-24 18:39:24,098 | step= 8940 | epoch=0.77 | loss=1.9117 | ppl=6.76 | lr=6.75e-05 | grad=0.292 | tok/s=268,678 | ETA=2:56:13 | VRAM=2.8GB | RAM=4%
2026-07-24 18:40:41,778 | step= 8960 | epoch=0.77 | loss=1.8057 | ppl=6.08 | lr=6.70e-05 | grad=0.273 | tok/s=268,837 | ETA=2:54:49 | VRAM=2.8GB | RAM=4%
2026-07-24 18:41:59,465 | step= 8980 | epoch=0.77 | loss=1.8845 | ppl=6.58 | lr=6.65e-05 | grad=0.269 | tok/s=270,644 | ETA=2:52:22 | VRAM=2.8GB | RAM=4%
2026-07-24 18:43:17,141 | step= 9000 | epoch=0.77 | loss=1.8566 | ppl=6.40 | lr=6.60e-05 | grad=0.306 | tok/s=269,440 | ETA=2:51:50 | VRAM=2.8GB | RAM=4%
2026-07-24 18:44:34,952 | step= 9020 | epoch=0.77 | loss=1.8638 | ppl=6.45 | lr=6.55e-05 | grad=0.285 | tok/s=269,912 | ETA=2:50:15 | VRAM=2.8GB | RAM=4%
2026-07-24 18:45:52,810 | step= 9040 | epoch=0.78 | loss=1.8424 | ppl=6.31 | lr=6.50e-05 | grad=0.271 | tok/s=269,519 | ETA=2:49:12 | VRAM=2.8GB | RAM=4%
2026-07-24 18:47:10,732 | step= 9060 | epoch=0.78 | loss=1.9069 | ppl=6.73 | lr=6.44e-05 | grad=0.311 | tok/s=268,620 | ETA=2:48:28 | VRAM=2.8GB | RAM=4%
2026-07-24 18:48:28,650 | step= 9080 | epoch=0.78 | loss=1.8425 | ppl=6.31 | lr=6.39e-05 | grad=0.301 | tok/s=268,382 | ETA=2:47:19 | VRAM=2.8GB | RAM=4%
2026-07-24 18:49:46,630 | step= 9100 | epoch=0.78 | loss=1.8666 | ppl=6.47 | lr=6.34e-05 | grad=0.282 | tok/s=269,944 | ETA=2:45:03 | VRAM=4.9GB | RAM=4%
2026-07-24 18:51:04,352 | step= 9120 | epoch=0.78 | loss=1.8397 | ppl=6.29 | lr=6.29e-05 | grad=0.285 | tok/s=269,876 | ETA=2:43:48 | VRAM=4.9GB | RAM=4%
2026-07-24 18:52:22,151 | step= 9140 | epoch=0.78 | loss=1.8387 | ppl=6.29 | lr=6.24e-05 | grad=0.285 | tok/s=268,713 | ETA=2:43:12 | VRAM=4.9GB | RAM=4%
2026-07-24 18:53:40,059 | step= 9160 | epoch=0.79 | loss=1.8607 | ppl=6.43 | lr=6.19e-05 | grad=0.271 | tok/s=269,246 | ETA=2:41:35 | VRAM=4.9GB | RAM=4%
2026-07-24 18:54:57,843 | step= 9180 | epoch=0.79 | loss=1.8273 | ppl=6.22 | lr=6.15e-05 | grad=0.286 | tok/s=267,633 | ETA=2:41:15 | VRAM=4.9GB | RAM=4%
2026-07-24 18:56:15,662 | step= 9200 | epoch=0.79 | loss=1.8070 | ppl=6.09 | lr=6.10e-05 | grad=0.280 | tok/s=269,847 | ETA=2:38:38 | VRAM=4.9GB | RAM=4%
2026-07-24 19:01:59,447 | VAL step=9200 epoch=0.79 loss=1.9101 ppl=6.75 ★ BEST
2026-07-24 19:02:00,045 | Checkpoint saved: step 9200
2026-07-24 19:02:00,465 | Saved safetensors: 128 tensors
2026-07-24 19:02:01,368 | HTTP Request: POST https://huggingface.co/api/repos/create "HTTP/1.1 409 Conflict"
2026-07-24 19:02:01,389 | HTTP Request: POST https://huggingface.co/api/validate-yaml "HTTP/1.1 200 OK"
2026-07-24 19:02:01,451 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/preupload/main "HTTP/1.1 200 OK"
2026-07-24 19:02:08,889 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/commit/main "HTTP/1.1 200 OK"
2026-07-24 19:02:08,896 | Hub push step=9200
2026-07-24 19:03:26,784 | step= 9220 | epoch=0.79 | loss=1.8407 | ppl=6.30 | lr=6.05e-05 | grad=0.288 | tok/s=269,275 | ETA=2:37:41 | VRAM=2.8GB | RAM=4%
2026-07-24 19:04:44,616 | step= 9240 | epoch=0.79 | loss=1.8050 | ppl=6.08 | lr=6.00e-05 | grad=0.288 | tok/s=269,726 | ETA=2:36:07 | VRAM=2.8GB | RAM=4%
2026-07-24 19:06:02,318 | step= 9260 | epoch=0.79 | loss=1.8504 | ppl=6.36 | lr=5.95e-05 | grad=0.316 | tok/s=269,143 | ETA=2:35:10 | VRAM=2.8GB | RAM=4%
2026-07-24 19:07:20,116 | step= 9280 | epoch=0.80 | loss=1.7673 | ppl=5.86 | lr=5.91e-05 | grad=0.295 | tok/s=270,519 | ETA=2:33:05 | VRAM=2.8GB | RAM=4%
2026-07-24 19:08:37,936 | step= 9300 | epoch=0.80 | loss=1.8447 | ppl=6.33 | lr=5.86e-05 | grad=0.277 | tok/s=268,534 | ETA=2:32:55 | VRAM=2.8GB | RAM=4%
2026-07-24 19:09:55,697 | step= 9320 | epoch=0.80 | loss=1.8060 | ppl=6.09 | lr=5.81e-05 | grad=0.278 | tok/s=269,601 | ETA=2:31:01 | VRAM=2.8GB | RAM=4%
2026-07-24 19:11:13,391 | step= 9340 | epoch=0.80 | loss=1.8835 | ppl=6.58 | lr=5.77e-05 | grad=0.282 | tok/s=270,046 | ETA=2:29:28 | VRAM=2.8GB | RAM=4%
2026-07-24 19:12:31,286 | step= 9360 | epoch=0.80 | loss=1.8209 | ppl=6.18 | lr=5.72e-05 | grad=0.281 | tok/s=268,722 | ETA=2:28:54 | VRAM=2.8GB | RAM=4%
2026-07-24 19:13:49,046 | step= 9380 | epoch=0.81 | loss=1.8268 | ppl=6.21 | lr=5.67e-05 | grad=0.288 | tok/s=269,370 | ETA=2:27:15 | VRAM=2.8GB | RAM=4%
2026-07-24 19:15:06,849 | step= 9400 | epoch=0.81 | loss=1.8596 | ppl=6.42 | lr=5.63e-05 | grad=0.333 | tok/s=270,002 | ETA=2:25:37 | VRAM=2.8GB | RAM=4%
2026-07-24 19:16:24,874 | step= 9420 | epoch=0.81 | loss=1.7850 | ppl=5.96 | lr=5.58e-05 | grad=0.281 | tok/s=268,378 | ETA=2:25:12 | VRAM=2.8GB | RAM=4%
2026-07-24 19:17:42,627 | step= 9440 | epoch=0.81 | loss=1.8457 | ppl=6.33 | lr=5.54e-05 | grad=0.289 | tok/s=270,474 | ETA=2:22:47 | VRAM=2.8GB | RAM=4%
2026-07-24 19:19:00,456 | step= 9460 | epoch=0.81 | loss=1.8638 | ppl=6.45 | lr=5.50e-05 | grad=0.309 | tok/s=269,360 | ETA=2:22:04 | VRAM=2.8GB | RAM=4%
2026-07-24 19:20:18,261 | step= 9480 | epoch=0.81 | loss=1.8521 | ppl=6.37 | lr=5.45e-05 | grad=0.287 | tok/s=269,680 | ETA=2:20:36 | VRAM=2.8GB | RAM=4%
2026-07-24 19:21:36,039 | step= 9500 | epoch=0.82 | loss=1.7662 | ppl=5.85 | lr=5.41e-05 | grad=0.306 | tok/s=268,422 | ETA=2:19:58 | VRAM=2.8GB | RAM=4%
2026-07-24 19:22:53,884 | step= 9520 | epoch=0.82 | loss=1.8966 | ppl=6.66 | lr=5.36e-05 | grad=0.301 | tok/s=269,291 | ETA=2:18:13 | VRAM=2.8GB | RAM=4%
2026-07-24 19:24:11,644 | step= 9540 | epoch=0.82 | loss=1.8071 | ppl=6.09 | lr=5.32e-05 | grad=0.289 | tok/s=269,399 | ETA=2:16:52 | VRAM=2.8GB | RAM=4%
2026-07-24 19:25:29,522 | step= 9560 | epoch=0.82 | loss=1.8315 | ppl=6.24 | lr=5.28e-05 | grad=0.298 | tok/s=269,282 | ETA=2:15:38 | VRAM=2.8GB | RAM=4%
2026-07-24 19:26:47,285 | step= 9580 | epoch=0.82 | loss=1.8733 | ppl=6.51 | lr=5.24e-05 | grad=0.305 | tok/s=269,561 | ETA=2:14:12 | VRAM=2.8GB | RAM=4%
2026-07-24 19:28:05,156 | step= 9600 | epoch=0.82 | loss=1.8324 | ppl=6.25 | lr=5.20e-05 | grad=0.265 | tok/s=269,896 | ETA=2:12:44 | VRAM=2.8GB | RAM=4%
2026-07-24 19:33:14,735 | VAL step=9600 epoch=0.82 loss=1.9050 ppl=6.72 ★ BEST
2026-07-24 19:33:15,318 | Checkpoint saved: step 9600
2026-07-24 19:33:15,735 | Saved safetensors: 128 tensors
2026-07-24 19:33:16,442 | HTTP Request: POST https://huggingface.co/api/repos/create "HTTP/1.1 409 Conflict"
2026-07-24 19:33:16,464 | HTTP Request: POST https://huggingface.co/api/validate-yaml "HTTP/1.1 200 OK"
2026-07-24 19:33:16,530 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/preupload/main "HTTP/1.1 200 OK"
2026-07-24 19:33:23,519 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/commit/main "HTTP/1.1 200 OK"
2026-07-24 19:33:23,525 | Hub push step=9600
2026-07-24 19:34:41,374 | step= 9620 | epoch=0.83 | loss=1.8528 | ppl=6.38 | lr=5.15e-05 | grad=0.315 | tok/s=270,132 | ETA=2:11:19 | VRAM=2.8GB | RAM=4%
2026-07-24 19:35:59,102 | step= 9640 | epoch=0.83 | loss=1.8665 | ppl=6.47 | lr=5.11e-05 | grad=0.293 | tok/s=269,662 | ETA=2:10:15 | VRAM=2.8GB | RAM=4%
2026-07-24 19:37:16,910 | step= 9660 | epoch=0.83 | loss=1.8581 | ppl=6.41 | lr=5.07e-05 | grad=0.285 | tok/s=268,408 | ETA=2:09:34 | VRAM=2.8GB | RAM=4%
2026-07-24 19:38:34,714 | step= 9680 | epoch=0.83 | loss=1.8348 | ppl=6.26 | lr=5.03e-05 | grad=0.268 | tok/s=270,137 | ETA=2:07:27 | VRAM=2.8GB | RAM=4%
2026-07-24 19:39:52,534 | step= 9700 | epoch=0.83 | loss=1.9124 | ppl=6.77 | lr=4.99e-05 | grad=0.305 | tok/s=268,228 | ETA=2:07:03 | VRAM=2.8GB | RAM=4%
2026-07-24 19:41:10,225 | step= 9720 | epoch=0.83 | loss=1.8898 | ppl=6.62 | lr=4.95e-05 | grad=0.298 | tok/s=270,479 | ETA=2:04:42 | VRAM=2.8GB | RAM=4%
2026-07-24 19:42:28,011 | step= 9740 | epoch=0.84 | loss=1.8587 | ppl=6.42 | lr=4.91e-05 | grad=0.290 | tok/s=269,697 | ETA=2:03:46 | VRAM=2.8GB | RAM=4%
2026-07-24 19:43:45,826 | step= 9760 | epoch=0.84 | loss=1.8857 | ppl=6.59 | lr=4.87e-05 | grad=0.277 | tok/s=269,155 | ETA=2:02:43 | VRAM=2.8GB | RAM=4%
2026-07-24 19:45:03,615 | step= 9780 | epoch=0.84 | loss=1.8397 | ppl=6.29 | lr=4.84e-05 | grad=0.313 | tok/s=268,769 | ETA=2:01:36 | VRAM=2.8GB | RAM=4%
2026-07-24 19:46:21,397 | step= 9800 | epoch=0.84 | loss=1.8463 | ppl=6.34 | lr=4.80e-05 | grad=0.269 | tok/s=269,046 | ETA=2:00:10 | VRAM=2.8GB | RAM=4%
2026-07-24 19:47:39,200 | step= 9820 | epoch=0.84 | loss=1.8539 | ppl=6.38 | lr=4.76e-05 | grad=0.286 | tok/s=269,634 | ETA=1:58:37 | VRAM=2.8GB | RAM=4%
2026-07-24 19:48:56,937 | step= 9840 | epoch=0.84 | loss=1.8372 | ppl=6.28 | lr=4.72e-05 | grad=0.265 | tok/s=269,518 | ETA=1:57:22 | VRAM=2.8GB | RAM=4%
2026-07-24 19:50:14,645 | step= 9860 | epoch=0.85 | loss=1.8200 | ppl=6.17 | lr=4.69e-05 | grad=0.329 | tok/s=268,917 | ETA=1:56:20 | VRAM=2.8GB | RAM=4%
2026-07-24 19:51:32,400 | step= 9880 | epoch=0.85 | loss=1.8367 | ppl=6.28 | lr=4.65e-05 | grad=0.299 | tok/s=268,454 | ETA=1:55:14 | VRAM=2.8GB | RAM=4%
2026-07-24 19:52:50,183 | step= 9900 | epoch=0.85 | loss=1.8470 | ppl=6.34 | lr=4.61e-05 | grad=0.269 | tok/s=268,964 | ETA=1:53:43 | VRAM=2.8GB | RAM=4%
2026-07-24 19:54:07,889 | step= 9920 | epoch=0.85 | loss=1.8619 | ppl=6.44 | lr=4.58e-05 | grad=0.292 | tok/s=270,258 | ETA=1:51:53 | VRAM=2.8GB | RAM=4%
2026-07-24 19:55:25,661 | step= 9940 | epoch=0.85 | loss=1.8531 | ppl=6.38 | lr=4.54e-05 | grad=0.303 | tok/s=270,313 | ETA=1:50:34 | VRAM=2.8GB | RAM=4%
2026-07-24 19:56:44,022 | step= 9960 | epoch=0.85 | loss=1.8162 | ppl=6.15 | lr=4.51e-05 | grad=0.271 | tok/s=268,822 | ETA=1:49:53 | VRAM=2.8GB | RAM=4%
2026-07-24 19:58:02,274 | step= 9980 | epoch=0.86 | loss=1.8523 | ppl=6.37 | lr=4.47e-05 | grad=0.302 | tok/s=268,224 | ETA=1:48:50 | VRAM=2.8GB | RAM=4%
2026-07-24 19:59:20,467 | step= 10000 | epoch=0.86 | loss=1.9057 | ppl=6.72 | lr=4.44e-05 | grad=0.271 | tok/s=268,494 | ETA=1:47:25 | VRAM=2.8GB | RAM=4%
2026-07-24 20:04:31,980 | VAL step=10000 epoch=0.86 loss=1.8998 ppl=6.68 ★ BEST
2026-07-24 20:04:32,558 | Checkpoint saved: step 10000
2026-07-24 20:04:32,780 | Saved safetensors: 128 tensors
2026-07-24 20:04:33,454 | HTTP Request: POST https://huggingface.co/api/repos/create "HTTP/1.1 409 Conflict"
2026-07-24 20:04:33,476 | HTTP Request: POST https://huggingface.co/api/validate-yaml "HTTP/1.1 200 OK"
2026-07-24 20:04:33,629 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/preupload/main "HTTP/1.1 200 OK"
2026-07-24 20:04:39,458 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/commit/main "HTTP/1.1 200 OK"
2026-07-24 20:04:39,463 | Hub push step=10000
2026-07-24 20:05:57,622 | step= 10020 | epoch=0.86 | loss=1.8491 | ppl=6.35 | lr=4.40e-05 | grad=0.268 | tok/s=268,508 | ETA=1:46:07 | VRAM=2.8GB | RAM=4%
2026-07-24 20:07:15,369 | step= 10040 | epoch=0.86 | loss=1.8440 | ppl=6.32 | lr=4.37e-05 | grad=0.255 | tok/s=270,180 | ETA=1:44:10 | VRAM=2.8GB | RAM=4%
2026-07-24 20:08:32,997 | step= 10060 | epoch=0.86 | loss=1.8688 | ppl=6.48 | lr=4.34e-05 | grad=0.280 | tok/s=270,413 | ETA=1:42:47 | VRAM=2.8GB | RAM=4%
2026-07-24 20:09:50,873 | step= 10080 | epoch=0.87 | loss=1.8438 | ppl=6.32 | lr=4.30e-05 | grad=0.311 | tok/s=268,577 | ETA=1:42:11 | VRAM=2.8GB | RAM=4%
2026-07-24 20:11:08,709 | step= 10100 | epoch=0.87 | loss=1.9906 | ppl=7.32 | lr=4.27e-05 | grad=0.380 | tok/s=269,403 | ETA=1:40:34 | VRAM=2.8GB | RAM=4%
2026-07-24 20:12:26,669 | step= 10120 | epoch=0.87 | loss=1.7548 | ppl=5.78 | lr=4.24e-05 | grad=0.281 | tok/s=267,511 | ETA=1:39:59 | VRAM=2.8GB | RAM=4%
2026-07-24 20:13:44,797 | step= 10140 | epoch=0.87 | loss=1.9957 | ppl=7.36 | lr=4.21e-05 | grad=0.836 | tok/s=268,305 | ETA=1:38:23 | VRAM=2.8GB | RAM=4%
2026-07-24 20:15:02,766 | step= 10160 | epoch=0.87 | loss=1.8697 | ppl=6.49 | lr=4.18e-05 | grad=0.309 | tok/s=268,553 | ETA=1:36:59 | VRAM=2.8GB | RAM=4%
2026-07-24 20:16:20,661 | step= 10180 | epoch=0.87 | loss=1.8329 | ppl=6.25 | lr=4.15e-05 | grad=0.259 | tok/s=269,278 | ETA=1:35:26 | VRAM=2.8GB | RAM=4%
2026-07-24 20:17:38,607 | step= 10200 | epoch=0.88 | loss=1.8581 | ppl=6.41 | lr=4.12e-05 | grad=0.302 | tok/s=267,794 | ETA=1:34:39 | VRAM=2.8GB | RAM=4%
2026-07-24 20:18:56,652 | step= 10220 | epoch=0.88 | loss=1.8047 | ppl=6.08 | lr=4.09e-05 | grad=0.293 | tok/s=266,895 | ETA=1:33:40 | VRAM=2.8GB | RAM=4%
2026-07-24 20:20:14,962 | step= 10240 | epoch=0.88 | loss=1.7806 | ppl=5.93 | lr=4.06e-05 | grad=0.267 | tok/s=266,389 | ETA=1:32:32 | VRAM=2.8GB | RAM=4%
2026-07-24 20:21:33,422 | step= 10260 | epoch=0.88 | loss=1.7963 | ppl=6.03 | lr=4.03e-05 | grad=0.287 | tok/s=267,578 | ETA=1:30:49 | VRAM=2.8GB | RAM=4%
2026-07-24 20:22:51,726 | step= 10280 | epoch=0.88 | loss=1.8656 | ppl=6.46 | lr=4.00e-05 | grad=0.304 | tok/s=267,173 | ETA=1:29:39 | VRAM=2.8GB | RAM=4%
2026-07-24 20:24:10,009 | step= 10300 | epoch=0.88 | loss=1.7857 | ppl=5.96 | lr=3.97e-05 | grad=0.290 | tok/s=269,065 | ETA=1:27:43 | VRAM=2.8GB | RAM=4%
2026-07-24 20:25:28,035 | step= 10320 | epoch=0.89 | loss=1.8017 | ppl=6.06 | lr=3.94e-05 | grad=0.276 | tok/s=268,776 | ETA=1:26:31 | VRAM=2.8GB | RAM=4%
2026-07-24 20:26:46,084 | step= 10340 | epoch=0.89 | loss=1.8948 | ppl=6.65 | lr=3.91e-05 | grad=0.281 | tok/s=269,708 | ETA=1:24:55 | VRAM=2.8GB | RAM=4%
2026-07-24 20:28:04,051 | step= 10360 | epoch=0.89 | loss=1.8680 | ppl=6.48 | lr=3.89e-05 | grad=0.289 | tok/s=266,907 | ETA=1:24:30 | VRAM=2.8GB | RAM=4%
2026-07-24 20:29:21,966 | step= 10380 | epoch=0.89 | loss=1.8093 | ppl=6.11 | lr=3.86e-05 | grad=0.276 | tok/s=268,032 | ETA=1:22:51 | VRAM=2.8GB | RAM=4%
2026-07-24 20:30:39,764 | step= 10400 | epoch=0.89 | loss=1.8634 | ppl=6.45 | lr=3.83e-05 | grad=0.276 | tok/s=269,515 | ETA=1:21:06 | VRAM=2.8GB | RAM=4%
2026-07-24 20:35:51,744 | VAL step=10400 epoch=0.89 loss=1.8960 ppl=6.66 ★ BEST
2026-07-24 20:35:52,391 | Checkpoint saved: step 10400
2026-07-24 20:35:52,629 | Saved safetensors: 128 tensors
2026-07-24 20:35:53,596 | HTTP Request: POST https://huggingface.co/api/repos/create "HTTP/1.1 409 Conflict"
2026-07-24 20:35:53,619 | HTTP Request: POST https://huggingface.co/api/validate-yaml "HTTP/1.1 200 OK"
2026-07-24 20:35:53,673 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/preupload/main "HTTP/1.1 200 OK"
2026-07-24 20:36:01,183 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/commit/main "HTTP/1.1 200 OK"
2026-07-24 20:36:01,190 | Hub push step=10400
2026-07-24 20:37:19,185 | step= 10420 | epoch=0.89 | loss=1.8367 | ppl=6.28 | lr=3.81e-05 | grad=0.294 | tok/s=269,255 | ETA=1:19:53 | VRAM=2.8GB | RAM=4%
2026-07-24 20:38:37,284 | step= 10440 | epoch=0.90 | loss=1.8793 | ppl=6.55 | lr=3.78e-05 | grad=0.270 | tok/s=268,954 | ETA=1:18:40 | VRAM=2.8GB | RAM=4%
2026-07-24 20:39:55,459 | step= 10460 | epoch=0.90 | loss=1.8390 | ppl=6.29 | lr=3.76e-05 | grad=0.274 | tok/s=267,856 | ETA=1:17:41 | VRAM=2.8GB | RAM=4%
2026-07-24 20:41:13,643 | step= 10480 | epoch=0.90 | loss=1.8709 | ppl=6.49 | lr=3.73e-05 | grad=0.289 | tok/s=268,423 | ETA=1:16:13 | VRAM=2.8GB | RAM=4%
2026-07-24 20:42:31,742 | step= 10500 | epoch=0.90 | loss=1.8841 | ppl=6.58 | lr=3.71e-05 | grad=0.278 | tok/s=266,778 | ETA=1:15:23 | VRAM=2.8GB | RAM=4%
2026-07-24 20:43:49,784 | step= 10520 | epoch=0.90 | loss=1.9385 | ppl=6.95 | lr=3.68e-05 | grad=0.425 | tok/s=268,819 | ETA=1:13:31 | VRAM=2.8GB | RAM=4%
2026-07-24 20:45:07,701 | step= 10540 | epoch=0.90 | loss=1.8205 | ppl=6.18 | lr=3.66e-05 | grad=0.310 | tok/s=266,670 | ETA=1:12:48 | VRAM=2.8GB | RAM=4%
2026-07-24 20:46:25,767 | step= 10560 | epoch=0.91 | loss=1.8730 | ppl=6.51 | lr=3.63e-05 | grad=0.321 | tok/s=268,010 | ETA=1:11:08 | VRAM=2.8GB | RAM=4%
2026-07-24 20:47:43,699 | step= 10580 | epoch=0.91 | loss=1.8236 | ppl=6.19 | lr=3.61e-05 | grad=0.302 | tok/s=269,219 | ETA=1:09:31 | VRAM=2.8GB | RAM=4%
2026-07-24 20:49:01,598 | step= 10600 | epoch=0.91 | loss=1.7434 | ppl=5.72 | lr=3.59e-05 | grad=0.330 | tok/s=269,249 | ETA=1:08:12 | VRAM=2.8GB | RAM=4%
2026-07-24 20:50:19,813 | step= 10620 | epoch=0.91 | loss=1.8435 | ppl=6.32 | lr=3.57e-05 | grad=0.302 | tok/s=268,201 | ETA=1:07:10 | VRAM=2.8GB | RAM=4%
2026-07-24 20:51:37,783 | step= 10640 | epoch=0.91 | loss=1.8317 | ppl=6.24 | lr=3.55e-05 | grad=0.314 | tok/s=268,062 | ETA=1:05:54 | VRAM=2.8GB | RAM=4%
2026-07-24 20:52:55,840 | step= 10660 | epoch=0.91 | loss=1.8605 | ppl=6.43 | lr=3.52e-05 | grad=0.273 | tok/s=267,554 | ETA=1:04:43 | VRAM=2.8GB | RAM=4%
2026-07-24 20:54:13,923 | step= 10680 | epoch=0.92 | loss=1.8194 | ppl=6.17 | lr=3.50e-05 | grad=0.324 | tok/s=269,258 | ETA=1:03:01 | VRAM=2.8GB | RAM=4%
2026-07-24 20:55:32,049 | step= 10700 | epoch=0.92 | loss=1.8464 | ppl=6.34 | lr=3.48e-05 | grad=0.334 | tok/s=268,389 | ETA=1:01:55 | VRAM=2.8GB | RAM=4%
2026-07-24 20:56:50,129 | step= 10720 | epoch=0.92 | loss=1.8211 | ppl=6.18 | lr=3.46e-05 | grad=0.320 | tok/s=268,854 | ETA=1:00:31 | VRAM=2.8GB | RAM=4%
2026-07-24 20:58:07,995 | step= 10740 | epoch=0.92 | loss=1.7889 | ppl=5.98 | lr=3.44e-05 | grad=0.353 | tok/s=269,902 | ETA=0:58:59 | VRAM=2.8GB | RAM=4%
2026-07-24 20:59:26,042 | step= 10760 | epoch=0.92 | loss=1.8632 | ppl=6.44 | lr=3.42e-05 | grad=0.268 | tok/s=269,196 | ETA=0:57:51 | VRAM=2.8GB | RAM=4%
2026-07-24 21:00:43,992 | step= 10780 | epoch=0.93 | loss=1.9256 | ppl=6.86 | lr=3.41e-05 | grad=0.261 | tok/s=268,817 | ETA=0:56:38 | VRAM=2.8GB | RAM=4%
2026-07-24 21:02:01,953 | step= 10800 | epoch=0.93 | loss=1.8872 | ppl=6.60 | lr=3.39e-05 | grad=0.249 | tok/s=268,131 | ETA=0:55:28 | VRAM=2.8GB | RAM=4%
2026-07-24 21:07:14,893 | VAL step=10800 epoch=0.93 loss=1.8945 ppl=6.65 ★ BEST
2026-07-24 21:07:15,503 | Checkpoint saved: step 10800
2026-07-24 21:07:16,115 | Saved safetensors: 128 tensors
2026-07-24 21:07:16,834 | HTTP Request: POST https://huggingface.co/api/repos/create "HTTP/1.1 409 Conflict"
2026-07-24 21:07:16,862 | HTTP Request: POST https://huggingface.co/api/validate-yaml "HTTP/1.1 200 OK"
2026-07-24 21:07:16,910 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/preupload/main "HTTP/1.1 200 OK"
2026-07-24 21:07:22,969 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/commit/main "HTTP/1.1 200 OK"
2026-07-24 21:07:22,977 | Hub push step=10800
2026-07-24 21:08:40,856 | step= 10820 | epoch=0.93 | loss=1.8496 | ppl=6.36 | lr=3.37e-05 | grad=0.285 | tok/s=269,643 | ETA=0:53:52 | VRAM=2.8GB | RAM=4%
2026-07-24 21:09:58,650 | step= 10840 | epoch=0.93 | loss=1.9015 | ppl=6.70 | lr=3.35e-05 | grad=0.265 | tok/s=267,937 | ETA=0:52:54 | VRAM=2.8GB | RAM=4%
2026-07-24 21:11:16,399 | step= 10860 | epoch=0.93 | loss=1.8430 | ppl=6.32 | lr=3.34e-05 | grad=0.270 | tok/s=268,045 | ETA=0:51:35 | VRAM=2.8GB | RAM=4%
2026-07-24 21:12:34,169 | step= 10880 | epoch=0.93 | loss=1.8847 | ppl=6.58 | lr=3.32e-05 | grad=0.282 | tok/s=268,891 | ETA=0:50:07 | VRAM=2.8GB | RAM=4%
2026-07-24 21:13:51,867 | step= 10900 | epoch=0.94 | loss=1.8696 | ppl=6.49 | lr=3.30e-05 | grad=0.274 | tok/s=269,174 | ETA=0:48:46 | VRAM=2.8GB | RAM=4%
2026-07-24 21:15:09,741 | step= 10920 | epoch=0.94 | loss=1.8623 | ppl=6.44 | lr=3.29e-05 | grad=0.269 | tok/s=268,914 | ETA=0:47:31 | VRAM=2.8GB | RAM=4%
2026-07-24 21:16:27,635 | step= 10940 | epoch=0.94 | loss=1.8284 | ppl=6.22 | lr=3.27e-05 | grad=0.266 | tok/s=268,764 | ETA=0:46:15 | VRAM=2.8GB | RAM=4%
2026-07-24 21:17:45,480 | step= 10960 | epoch=0.94 | loss=1.8252 | ppl=6.20 | lr=3.26e-05 | grad=0.257 | tok/s=269,449 | ETA=0:44:50 | VRAM=2.8GB | RAM=4%
2026-07-24 21:19:03,274 | step= 10980 | epoch=0.94 | loss=1.7898 | ppl=5.99 | lr=3.24e-05 | grad=0.270 | tok/s=269,022 | ETA=0:43:36 | VRAM=2.8GB | RAM=4%
2026-07-24 21:20:21,024 | step= 11000 | epoch=0.94 | loss=1.8794 | ppl=6.55 | lr=3.23e-05 | grad=0.286 | tok/s=269,528 | ETA=0:42:14 | VRAM=2.8GB | RAM=4%
2026-07-24 21:21:38,871 | step= 11020 | epoch=0.95 | loss=1.8688 | ppl=6.48 | lr=3.21e-05 | grad=0.289 | tok/s=269,424 | ETA=0:40:57 | VRAM=2.8GB | RAM=4%
2026-07-24 21:22:56,606 | step= 11040 | epoch=0.95 | loss=1.7705 | ppl=5.87 | lr=3.20e-05 | grad=0.276 | tok/s=268,932 | ETA=0:39:43 | VRAM=2.8GB | RAM=4%
2026-07-24 21:24:14,422 | step= 11060 | epoch=0.95 | loss=1.8436 | ppl=6.32 | lr=3.19e-05 | grad=0.284 | tok/s=268,462 | ETA=0:38:30 | VRAM=2.8GB | RAM=4%
2026-07-24 21:25:32,216 | step= 11080 | epoch=0.95 | loss=1.8325 | ppl=6.25 | lr=3.18e-05 | grad=0.259 | tok/s=269,722 | ETA=0:37:01 | VRAM=2.8GB | RAM=4%
2026-07-24 21:26:49,993 | step= 11100 | epoch=0.95 | loss=1.8142 | ppl=6.14 | lr=3.16e-05 | grad=0.291 | tok/s=268,992 | ETA=0:35:49 | VRAM=2.8GB | RAM=4%
2026-07-24 21:28:07,960 | step= 11120 | epoch=0.95 | loss=1.8339 | ppl=6.26 | lr=3.15e-05 | grad=0.290 | tok/s=269,166 | ETA=0:34:30 | VRAM=2.8GB | RAM=4%
2026-07-24 21:29:25,829 | step= 11140 | epoch=0.96 | loss=1.8483 | ppl=6.35 | lr=3.14e-05 | grad=0.295 | tok/s=268,703 | ETA=0:33:16 | VRAM=2.8GB | RAM=4%
2026-07-24 21:30:43,692 | step= 11160 | epoch=0.96 | loss=1.8846 | ppl=6.58 | lr=3.13e-05 | grad=0.264 | tok/s=268,960 | ETA=0:31:56 | VRAM=2.8GB | RAM=4%
2026-07-24 21:32:01,532 | step= 11180 | epoch=0.96 | loss=1.8764 | ppl=6.53 | lr=3.12e-05 | grad=0.270 | tok/s=269,321 | ETA=0:30:35 | VRAM=2.8GB | RAM=4%
2026-07-24 21:33:19,313 | step= 11200 | epoch=0.96 | loss=1.8890 | ppl=6.61 | lr=3.11e-05 | grad=0.263 | tok/s=269,877 | ETA=0:29:14 | VRAM=2.8GB | RAM=4%
2026-07-24 21:38:29,020 | VAL step=11200 epoch=0.96 loss=1.8925 ppl=6.64 ★ BEST
2026-07-24 21:38:29,613 | Checkpoint saved: step 11200
2026-07-24 21:38:29,843 | Saved safetensors: 128 tensors
2026-07-24 21:38:30,707 | HTTP Request: POST https://huggingface.co/api/repos/create "HTTP/1.1 409 Conflict"
2026-07-24 21:38:30,728 | HTTP Request: POST https://huggingface.co/api/validate-yaml "HTTP/1.1 200 OK"
2026-07-24 21:38:30,788 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/preupload/main "HTTP/1.1 200 OK"
2026-07-24 21:38:36,796 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/commit/main "HTTP/1.1 200 OK"
2026-07-24 21:38:36,800 | Hub push step=11200
2026-07-24 21:39:54,582 | step= 11220 | epoch=0.96 | loss=1.9293 | ppl=6.88 | lr=3.10e-05 | grad=0.296 | tok/s=268,539 | ETA=0:28:05 | VRAM=2.8GB | RAM=4%
2026-07-24 21:41:12,308 | step= 11240 | epoch=0.96 | loss=1.8908 | ppl=6.62 | lr=3.09e-05 | grad=0.261 | tok/s=269,335 | ETA=0:26:42 | VRAM=2.8GB | RAM=4%
2026-07-24 21:42:30,095 | step= 11260 | epoch=0.97 | loss=1.8585 | ppl=6.41 | lr=3.08e-05 | grad=0.298 | tok/s=268,240 | ETA=0:25:30 | VRAM=2.8GB | RAM=4%
2026-07-24 21:43:48,027 | step= 11280 | epoch=0.97 | loss=1.8414 | ppl=6.31 | lr=3.07e-05 | grad=0.281 | tok/s=268,501 | ETA=0:24:11 | VRAM=2.8GB | RAM=4%
2026-07-24 21:45:05,718 | step= 11300 | epoch=0.97 | loss=1.7130 | ppl=5.55 | lr=3.07e-05 | grad=0.302 | tok/s=268,684 | ETA=0:22:52 | VRAM=2.8GB | RAM=4%
2026-07-24 21:46:23,463 | step= 11320 | epoch=0.97 | loss=1.8483 | ppl=6.35 | lr=3.06e-05 | grad=0.305 | tok/s=270,240 | ETA=0:21:26 | VRAM=2.8GB | RAM=4%
2026-07-24 21:47:41,116 | step= 11340 | epoch=0.97 | loss=1.8507 | ppl=6.36 | lr=3.05e-05 | grad=0.270 | tok/s=269,813 | ETA=0:20:11 | VRAM=2.8GB | RAM=4%
2026-07-24 21:48:58,882 | step= 11360 | epoch=0.97 | loss=1.3970 | ppl=4.04 | lr=3.05e-05 | grad=0.303 | tok/s=268,805 | ETA=0:18:57 | VRAM=2.8GB | RAM=4%
2026-07-24 21:50:16,555 | step= 11380 | epoch=0.98 | loss=1.8791 | ppl=6.55 | lr=3.04e-05 | grad=0.317 | tok/s=270,385 | ETA=0:17:33 | VRAM=2.8GB | RAM=4%
2026-07-24 21:51:34,284 | step= 11400 | epoch=0.98 | loss=1.8049 | ppl=6.08 | lr=3.03e-05 | grad=0.280 | tok/s=270,251 | ETA=0:16:16 | VRAM=2.8GB | RAM=4%
2026-07-24 21:52:52,038 | step= 11420 | epoch=0.98 | loss=1.7863 | ppl=5.97 | lr=3.03e-05 | grad=0.287 | tok/s=269,264 | ETA=0:15:02 | VRAM=2.8GB | RAM=4%
2026-07-24 21:54:09,804 | step= 11440 | epoch=0.98 | loss=1.8108 | ppl=6.12 | lr=3.02e-05 | grad=0.289 | tok/s=269,881 | ETA=0:13:42 | VRAM=2.8GB | RAM=4%
2026-07-24 21:55:27,546 | step= 11460 | epoch=0.98 | loss=1.8683 | ppl=6.48 | lr=3.02e-05 | grad=0.307 | tok/s=268,847 | ETA=0:12:28 | VRAM=2.8GB | RAM=4%
2026-07-24 21:56:45,352 | step= 11480 | epoch=0.99 | loss=1.8130 | ppl=6.13 | lr=3.02e-05 | grad=0.326 | tok/s=268,924 | ETA=0:11:10 | VRAM=2.8GB | RAM=4%
2026-07-24 21:58:03,085 | step= 11500 | epoch=0.99 | loss=1.8736 | ppl=6.51 | lr=3.01e-05 | grad=0.292 | tok/s=269,398 | ETA=0:09:51 | VRAM=2.8GB | RAM=4%
2026-07-24 21:59:20,833 | step= 11520 | epoch=0.99 | loss=1.8192 | ppl=6.17 | lr=3.01e-05 | grad=0.261 | tok/s=269,717 | ETA=0:08:32 | VRAM=2.8GB | RAM=4%
2026-07-24 22:00:38,584 | step= 11540 | epoch=0.99 | loss=1.8387 | ppl=6.29 | lr=3.01e-05 | grad=0.267 | tok/s=268,893 | ETA=0:07:16 | VRAM=2.8GB | RAM=4%
2026-07-24 22:01:56,269 | step= 11560 | epoch=0.99 | loss=1.8156 | ppl=6.14 | lr=3.00e-05 | grad=0.332 | tok/s=269,900 | ETA=0:05:57 | VRAM=2.8GB | RAM=4%
2026-07-24 22:03:14,054 | step= 11580 | epoch=0.99 | loss=1.8363 | ppl=6.27 | lr=3.00e-05 | grad=0.289 | tok/s=269,549 | ETA=0:04:39 | VRAM=2.8GB | RAM=4%
2026-07-24 22:04:31,811 | step= 11600 | epoch=1.00 | loss=1.8210 | ppl=6.18 | lr=3.00e-05 | grad=0.277 | tok/s=270,289 | ETA=0:03:21 | VRAM=2.8GB | RAM=4%
2026-07-24 22:09:41,429 | VAL step=11600 epoch=1.00 loss=1.8918 ppl=6.63 ★ BEST
2026-07-24 22:09:42,057 | Checkpoint saved: step 11600
2026-07-24 22:09:42,263 | Saved safetensors: 128 tensors
2026-07-24 22:09:43,244 | HTTP Request: POST https://huggingface.co/api/repos/create "HTTP/1.1 409 Conflict"
2026-07-24 22:09:43,263 | HTTP Request: POST https://huggingface.co/api/validate-yaml "HTTP/1.1 200 OK"
2026-07-24 22:09:43,499 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/preupload/main "HTTP/1.1 200 OK"
2026-07-24 22:09:49,177 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/commit/main "HTTP/1.1 200 OK"
2026-07-24 22:09:49,182 | Hub push step=11600
2026-07-24 22:11:06,855 | step= 11620 | epoch=1.00 | loss=1.8464 | ppl=6.34 | lr=3.00e-05 | grad=0.338 | tok/s=270,730 | ETA=0:02:03 | VRAM=2.8GB | RAM=4%
2026-07-24 22:12:24,685 | step= 11640 | epoch=1.00 | loss=1.7985 | ppl=6.04 | lr=3.00e-05 | grad=0.308 | tok/s=268,458 | ETA=0:00:46 | VRAM=2.8GB | RAM=4%
2026-07-24 22:18:20,682 | Final eval loss=1.8920 ppl=6.63
2026-07-24 22:18:21,294 | Checkpoint saved: step 11652
2026-07-24 22:18:21,523 | Saved safetensors: 128 tensors
2026-07-24 22:18:22,225 | HTTP Request: POST https://huggingface.co/api/repos/create "HTTP/1.1 409 Conflict"
2026-07-24 22:18:22,244 | HTTP Request: POST https://huggingface.co/api/validate-yaml "HTTP/1.1 200 OK"
2026-07-24 22:18:22,572 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/preupload/main "HTTP/1.1 200 OK"
2026-07-24 22:18:28,765 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/commit/main "HTTP/1.1 200 OK"
2026-07-24 22:18:28,770 | Hub push step=11652
2026-07-24 22:18:28,849 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/preupload/main "HTTP/1.1 200 OK"
2026-07-24 22:18:29,829 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/commit/main "HTTP/1.1 200 OK"
2026-07-24 22:18:29,892 | HTTP Request: POST https://huggingface.co/api/models/CyanMonkey/Petal-50M/preupload/main "HTTP/1.1 200 OK"

8
training_meta.json Normal file
View File

@@ -0,0 +1,8 @@
{
"step": 11652,
"val_loss": 1.892010307149887,
"val_ppl": 6.632689029740431,
"params_M": 49.691,
"pushed_at": "2026-07-24T22:18:21.538933",
"tokens_seen": 12206075904
}