134 lines
3.4 KiB
Markdown
134 lines
3.4 KiB
Markdown
---
|
|
library_name: transformers
|
|
base_model: Qwen/Qwen3-8B
|
|
tags:
|
|
- generated_from_trainer
|
|
# datasets: (stripped — axolotl embedded invalid local path)
|
|
_datasets_:
|
|
- laion/Sera-4.6-Lite-T2-v4-316
|
|
model-index:
|
|
- name: e/data1/datasets/playground/ot-baf/checkpoints/sera-v4-316-axolotl__Qwen3-8B-v3
|
|
results: []
|
|
---
|
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
should probably proofread and complete it, then remove this comment. -->
|
|
|
|
[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
|
|
<details><summary>See axolotl config</summary>
|
|
|
|
axolotl version: `0.16.0.dev0`
|
|
```yaml
|
|
# Sera v4 retrain v3 — size=316, num_epochs doubled 3->6
|
|
# v2 (num_epochs=3, tokenizer_default) recovered structure but model still
|
|
# emits malformed JSON ({"name": "view"}}} — 3 closing braces) and collapses
|
|
# str_replace_editor -> inner arguments.command='view'. Training data confirmed
|
|
# clean (5 rows / 140 tool calls / 0 malformed / tool names str_replace_editor/
|
|
# bash/submit). Hypothesis: ~120 grad updates at size=316 is too few to latch
|
|
# onto nested JSON structure + tool-name dictionary; double passes to 6 epochs.
|
|
|
|
base_model: Qwen/Qwen3-8B
|
|
deepspeed: /e/scratch/jureap59/feuer1/code/axolotl/deepspeed_configs/zero3_bf16.json
|
|
|
|
load_in_8bit: false
|
|
load_in_4bit: false
|
|
|
|
chat_template: tokenizer_default
|
|
# datasets: (stripped — axolotl embedded invalid local path)
|
|
_datasets_:
|
|
- laion/Sera-4.6-Lite-T2-v4-316
|
|
type: chat_template
|
|
field_messages: messages
|
|
ds_type: json
|
|
message_field_training: train
|
|
|
|
dataset_prepared_path: /e/data1/datasets/playground/ot-baf/axolotl_dataset_cache/sera-v4-316-v3
|
|
output_dir: /e/data1/datasets/playground/ot-baf/checkpoints/sera-v4-316-axolotl__Qwen3-8B-v3
|
|
|
|
sequence_len: 32768
|
|
|
|
wandb_project:
|
|
wandb_entity:
|
|
wandb_watch:
|
|
wandb_name: sera-v4-316-axolotl__Qwen3-8B-v3
|
|
wandb_log_model:
|
|
|
|
gradient_accumulation_steps: 8
|
|
micro_batch_size: 1
|
|
num_epochs: 6
|
|
optimizer: adamw_torch
|
|
lr_scheduler: cosine
|
|
learning_rate: 1e-5
|
|
adam_beta1: 0.9
|
|
adam_beta2: 0.95
|
|
|
|
bf16: auto
|
|
tf32: false
|
|
|
|
gradient_checkpointing: true
|
|
activation_offloading: true
|
|
resume_from_checkpoint:
|
|
logging_steps: 1
|
|
flash_attention: true
|
|
|
|
loss_watchdog_threshold: 5.0
|
|
loss_watchdog_patience: 3
|
|
|
|
warmup_ratio: 0.1875
|
|
evals_per_epoch: 0
|
|
save_strategy: epoch
|
|
|
|
weight_decay: 0.01
|
|
max_grad_norm: 1.0
|
|
special_tokens:
|
|
|
|
```
|
|
|
|
</details><br>
|
|
|
|
# e/data1/datasets/playground/ot-baf/checkpoints/sera-v4-316-axolotl__Qwen3-8B-v3
|
|
|
|
- laion/Sera-4.6-Lite-T2-v4-316
|
|
|
|
## Model description
|
|
|
|
More information needed
|
|
|
|
## Intended uses & limitations
|
|
|
|
More information needed
|
|
|
|
## Training and evaluation data
|
|
|
|
More information needed
|
|
|
|
## Training procedure
|
|
|
|
### Training hyperparameters
|
|
|
|
The following hyperparameters were used during training:
|
|
- learning_rate: 1e-05
|
|
- train_batch_size: 1
|
|
- eval_batch_size: 1
|
|
- seed: 42
|
|
- distributed_type: multi-GPU
|
|
- num_devices: 4
|
|
- gradient_accumulation_steps: 8
|
|
- total_train_batch_size: 32
|
|
- total_eval_batch_size: 4
|
|
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.95) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
|
|
- lr_scheduler_type: cosine
|
|
- lr_scheduler_warmup_steps: 6
|
|
- training_steps: 34
|
|
|
|
### Training results
|
|
|
|
|
|
|
|
### Framework versions
|
|
|
|
- Transformers 5.5.0
|
|
- Pytorch 2.9.1+cu130
|
|
- Datasets 4.5.0
|
|
- Tokenizers 0.22.2
|