初始化项目，由ModelHub XC社区提供模型

Model: lldois/v31_v29_live_specialist_r3_lr12e6_ep035 Source: Original Platform
2026-07-31 16:30:04 +08:00
commit 0e1452784b
17 changed files with 2572 additions and 0 deletions
--- a/.gitattributes
+++ b/.gitattributes
@@ -0,0 +1,36 @@
+*.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
+tokenizer.json filter=lfs diff=lfs merge=lfs -text
--- a/README.md
+++ b/README.md
@@ -0,0 +1,58 @@
+---
+library_name: transformers
+license: other
+base_model: lldois/v29_v19_user_world_guard_lr8e7_ep018
+tags:
+- llama-factory
+- full
+- generated_from_trainer
+model-index:
+- name: v31_v29_live_specialist_r3_lr12e6_ep035
+  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. -->
+
+# v31_v29_live_specialist_r3_lr12e6_ep035
+
+This model is a fine-tuned version of [lldois/v29_v19_user_world_guard_lr8e7_ep018](https://huggingface.co/lldois/v29_v19_user_world_guard_lr8e7_ep018) on the exp_v29_live_specialist_r3 dataset.
+
+## 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: 1.2e-06
+- train_batch_size: 1
+- eval_batch_size: 8
+- seed: 202607311
+- gradient_accumulation_steps: 4
+- total_train_batch_size: 4
+- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_steps: 0.02
+- num_epochs: 0.35
+
+### Training results
+
+
+
+### Framework versions
+
+- Transformers 5.6.0
+- Pytorch 2.7.1+cu126
+- Datasets 4.0.0
+- Tokenizers 0.22.2
--- a/all_results.json
+++ b/all_results.json
@@ -0,0 +1,8 @@
+{
+    "epoch": 0.3562043795620438,
+    "total_flos": 2.935955062939853e+16,
+    "train_loss": 1.5556521063945332,
+    "train_runtime": 514.4769,
+    "train_samples_per_second": 0.466,
+    "train_steps_per_second": 0.119
+}
--- a/chat_template.jinja
+++ b/chat_template.jinja
@@ -0,0 +1,89 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+    {%- set index = (messages|length - 1) - loop.index0 %}
+    {%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if message.content is string %}
+        {%- set content = message.content %}
+    {%- else %}
+        {%- set content = '' %}
+    {%- endif %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is string %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in content %}
+                {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+                {%- set content = content.split('</think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last or (not loop.last and reasoning_content) %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+    {%- if enable_thinking is defined and enable_thinking is false %}
+        {{- '<think>\n\n</think>\n\n' }}
+    {%- endif %}
+{%- endif %}
--- a/config.json
+++ b/config.json
@@ -0,0 +1,63 @@
+{
+  "architectures": [
+    "Qwen3ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": null,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 40960,
+  "max_window_layers": 28,
+  "model_type": "qwen3",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_parameters": {
+    "rope_theta": 1000000,
+    "rope_type": "default"
+  },
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "transformers_version": "5.6.0",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 176253
+}
--- a/experiment_recipe.json
+++ b/experiment_recipe.json
@@ -0,0 +1,68 @@
+{
+  "created_at": "2026-07-13 03:28:12",
+  "run": {
+    "name": "v31_v29_live_specialist_r3_lr12e6_ep035",
+    "dataset": "v29_live_specialist_r3",
+    "model_path": "/home/ll/llm4rec/experiments/outputs/v29_v19_user_world_guard_lr8e7_ep018",
+    "lr": "1.2e-6",
+    "epochs": 0.35,
+    "warmup": 0.02,
+    "scheduler": "cosine",
+    "seed": 202607311,
+    "note": "v29 continuation and live-domain specialist. Use diverse evidence-grounded three-level CoT plus strict no-think direct answers for every live-target record, with item/user/general-rec guards. Goal: raise recommendation-live without losing v29's speed or other dimensions.",
+    "config": "/home/ll/llm4rec/experiments/configs/v31_v29_live_specialist_r3_lr12e6_ep035.yaml"
+  },
+  "dataset_manifest": {
+    "name": "v29_live_specialist_r3",
+    "path": "/home/ll/llm4rec/experiments/data/v29_live_specialist_r3.jsonl",
+    "records": 13013,
+    "groups": {
+      "rec": 7413,
+      "user": 2400,
+      "item": 3200
+    },
+    "variants": {
+      "raw": 1271,
+      "user_strict_array": 1080,
+      "user_extra_no_think_logic": 424,
+      "rec_no_think_direct_final": 3071,
+      "item_no_think_direct_final": 1659,
+      "rec_r3_draft_evidence": 3071,
+      "item_short_think": 1541,
+      "user_strict_logic": 896
+    },
+    "sha256": "5710cee8621dde0b19e5418cfe81cc5aea2933d432478968c498ff0f00e24934",
+    "seed": 202607111
+  },
+  "dataset_recipe": "v29 live-domain specialist: all live-target raw CoT, individualized three-level R3 draft CoT, and no-think direct answers, plus balanced general-rec, item, and user guards.",
+  "cot_policy": "Preserve /think reasoning supervision and do not use v7_final_only as a CoT training base. For /no_think prompts, train pure final answers without generated <think> tags. This is route-specific behavior, not global CoT removal.",
+  "raw_counts": {
+    "rec": 19204,
+    "item": 10384,
+    "user": 2892
+  },
+  "eval_observations": {
+    "v07": "best local score so far: total=0.8978, eval_time≈47.3min; fast final outputs likely help.",
+    "v15": "best CoT-preserving score so far: total=0.8778, eval_time≈70.1min; logs show repeated tokens, JSON shell errors, prompt leakage, and verbose /no_think outputs.",
+    "v19": "best CoT-native continuation so far: total=0.8855, eval_time≈48.1min; user1 and rec4 improved but world dropped.",
+    "v20": "v7 final-only continuation with CoT restore failed as a CoT route: total=0.8527, item fell to 0.1840; do not use v7 as future CoT base.",
+    "v22": "scratch official-base 3 epoch clean CoT underperformed: total=0.8217; item/world preserved but user and rec2 are weak.",
+    "v23": "scratch official-base 5 epoch low-LR guard failed badly: total=0.6990; item/user collapse suggests long scratch SFT is not viable with current data mix.",
+    "v24": "v15 light repair is best among v22-v24 but still only total=0.8364; user2 improves but rec/world do not recover.",
+    "v25": "v19 product-heavy repair did not beat v19: total=0.8793. Logs show heavy repeated product tokens and repeated short-think phrases; avoid this over-sampling pattern.",
+    "v26": "best latest batch and fastest eval: total=0.8804, eval_time≈45.7min, best user1/rec1. It is useful as a base, but rec2/world dropped.",
+    "v27": "v12 product/ad repair kept user2/world relatively better but was slow and template-heavy: total=0.8563, eval_time≈70.6min. Do not continue this exact direction.",
+    "v28": "v26 no-template repair regressed to total=0.8748 and slowed to 51.7min; broad continuation from v26 did not recover world/rec balance.",
+    "v29": "new best CoT-native model: total=0.9038, eval_time≈45.3min. Strong item/user1/ad/product/world, with live recommendation (0.1054) the clearest remaining gap. Logs still show occasional no-think tag leakage and a repeated generic live-reasoning sentence.",
+    "v30": "three-source soup reached total=0.8846, below v29. It retained speed/world but diluted recommendation scores; do not repeat broad checkpoint averaging.",
+    "v16": "CoT pattern rewrite failed: total=0.7912; logs show malformed user JSON and fragmented recommendation reasoning.",
+    "v18": "low-LR mixed replay from v12 failed: total=0.8340; item/world dropped and rec outputs mixed text/itemic/think tags."
+  },
+  "script": "/home/ll/llm4rec/experiments/run_experiments.py",
+  "script_sha256": "a5a06949ec583ca8f6e5c6de8690d3005ec1ef801b508b7286f1162f5b0c87b4",
+  "deadline": "none",
+  "reproduce": {
+    "prepare_command": "EXPERIMENT_PREPARE_ONLY=1 python3 experiments/run_experiments.py",
+    "train_command": "CUDA_VISIBLE_DEVICES=<gpu> bash -lc 'source /home/ll/llm4rec/demo/LLaMA-Factory/.venv/bin/activate && llamafactory-cli train /home/ll/llm4rec/experiments/configs/v31_v29_live_specialist_r3_lr12e6_ep035.yaml'"
+  }
+}
--- a/generation_config.json
+++ b/generation_config.json
@@ -0,0 +1,12 @@
+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "temperature": 0.6,
+  "top_k": 20,
+  "top_p": 0.95,
+  "transformers_version": "5.6.0"
+}
--- a/model.safetensors
+++ b/model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0b772648f015a574a409246cadbaddfa1dd98de62da7b25434ea5490dae78fcc
+size 1602902832
--- a/runs/Jul13_03-28-40_ubuntu-System-Product-Name/events.out.tfevents.1783884520.ubuntu-System-Product-Name.1295876.0
+++ b/runs/Jul13_03-28-40_ubuntu-System-Product-Name/events.out.tfevents.1783884520.ubuntu-System-Product-Name.1295876.0
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6a56bcaa1a481ae033b41f025ac2d393d0735411bb84be980208b7873380d0c0
+size 8796
--- a/tokenizer.json
+++ b/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cd4d15f596979aecbc11ba12668cb21c9fb8452ce1235cd9d7878cc950170421
+size 16016554
--- a/tokenizer_config.json
+++ b/tokenizer_config.json
@@ -0,0 +1,16 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "is_local": true,
+  "local_files_only": false,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}
--- a/train.log
+++ b/train.log
--- a/train_results.json
+++ b/train_results.json
@@ -0,0 +1,8 @@
+{
+    "epoch": 0.3562043795620438,
+    "total_flos": 2.935955062939853e+16,
+    "train_loss": 1.5556521063945332,
+    "train_runtime": 514.4769,
+    "train_samples_per_second": 0.466,
+    "train_steps_per_second": 0.119
+}
--- a/trainer_log.jsonl
+++ b/trainer_log.jsonl
@@ -0,0 +1,13 @@
+{"current_steps": 5, "total_steps": 61, "loss": 1.613851547241211, "lr": 1.1966008801260823e-06, "epoch": 0.029197080291970802, "percentage": 8.2, "elapsed_time": "0:00:41", "remaining_time": "0:07:40"}
+{"current_steps": 10, "total_steps": 61, "loss": 1.3017424583435058, "lr": 1.1588017064714031e-06, "epoch": 0.058394160583941604, "percentage": 16.39, "elapsed_time": "0:01:24", "remaining_time": "0:07:11"}
+{"current_steps": 15, "total_steps": 61, "loss": 1.6337610244750977, "lr": 1.0816269827585781e-06, "epoch": 0.08759124087591241, "percentage": 24.59, "elapsed_time": "0:02:06", "remaining_time": "0:06:27"}
+{"current_steps": 20, "total_steps": 61, "loss": 1.509153175354004, "lr": 9.70514768967715e-07, "epoch": 0.11678832116788321, "percentage": 32.79, "elapsed_time": "0:02:47", "remaining_time": "0:05:43"}
+{"current_steps": 25, "total_steps": 61, "loss": 1.6203853607177734, "lr": 8.332945052839925e-07, "epoch": 0.145985401459854, "percentage": 40.98, "elapsed_time": "0:03:27", "remaining_time": "0:04:58"}
+{"current_steps": 30, "total_steps": 61, "loss": 1.6313507080078125, "lr": 6.796353163702331e-07, "epoch": 0.17518248175182483, "percentage": 49.18, "elapsed_time": "0:04:11", "remaining_time": "0:04:20"}
+{"current_steps": 35, "total_steps": 61, "loss": 1.5739914894104003, "lr": 5.203646836297668e-07, "epoch": 0.20437956204379562, "percentage": 57.38, "elapsed_time": "0:04:54", "remaining_time": "0:03:38"}
+{"current_steps": 40, "total_steps": 61, "loss": 1.5292200088500976, "lr": 3.6670549471600764e-07, "epoch": 0.23357664233576642, "percentage": 65.57, "elapsed_time": "0:05:38", "remaining_time": "0:02:57"}
+{"current_steps": 45, "total_steps": 61, "loss": 1.519069766998291, "lr": 2.294852310322849e-07, "epoch": 0.26277372262773724, "percentage": 73.77, "elapsed_time": "0:06:21", "remaining_time": "0:02:15"}
+{"current_steps": 50, "total_steps": 61, "loss": 1.5700279235839845, "lr": 1.1837301724142192e-07, "epoch": 0.291970802919708, "percentage": 81.97, "elapsed_time": "0:07:02", "remaining_time": "0:01:32"}
+{"current_steps": 55, "total_steps": 61, "loss": 1.6012556076049804, "lr": 4.119829352859694e-08, "epoch": 0.32116788321167883, "percentage": 90.16, "elapsed_time": "0:07:44", "remaining_time": "0:00:50"}
+{"current_steps": 60, "total_steps": 61, "loss": 1.5306562423706054, "lr": 3.399119873917744e-09, "epoch": 0.35036496350364965, "percentage": 98.36, "elapsed_time": "0:08:26", "remaining_time": "0:00:08"}
+{"current_steps": 61, "total_steps": 61, "epoch": 0.3562043795620438, "percentage": 100.0, "elapsed_time": "0:08:34", "remaining_time": "0:00:00"}
--- a/trainer_state.json
+++ b/trainer_state.json
@@ -0,0 +1,127 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.3562043795620438,
+  "eval_steps": 500,
+  "global_step": 61,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.029197080291970802,
+      "grad_norm": 9.75,
+      "learning_rate": 1.1966008801260823e-06,
+      "loss": 1.613851547241211,
+      "step": 5
+    },
+    {
+      "epoch": 0.058394160583941604,
+      "grad_norm": 7.6875,
+      "learning_rate": 1.1588017064714031e-06,
+      "loss": 1.3017424583435058,
+      "step": 10
+    },
+    {
+      "epoch": 0.08759124087591241,
+      "grad_norm": 8.5625,
+      "learning_rate": 1.0816269827585781e-06,
+      "loss": 1.6337610244750977,
+      "step": 15
+    },
+    {
+      "epoch": 0.11678832116788321,
+      "grad_norm": 7.625,
+      "learning_rate": 9.70514768967715e-07,
+      "loss": 1.509153175354004,
+      "step": 20
+    },
+    {
+      "epoch": 0.145985401459854,
+      "grad_norm": 7.59375,
+      "learning_rate": 8.332945052839925e-07,
+      "loss": 1.6203853607177734,
+      "step": 25
+    },
+    {
+      "epoch": 0.17518248175182483,
+      "grad_norm": 9.875,
+      "learning_rate": 6.796353163702331e-07,
+      "loss": 1.6313507080078125,
+      "step": 30
+    },
+    {
+      "epoch": 0.20437956204379562,
+      "grad_norm": 8.5,
+      "learning_rate": 5.203646836297668e-07,
+      "loss": 1.5739914894104003,
+      "step": 35
+    },
+    {
+      "epoch": 0.23357664233576642,
+      "grad_norm": 6.78125,
+      "learning_rate": 3.6670549471600764e-07,
+      "loss": 1.5292200088500976,
+      "step": 40
+    },
+    {
+      "epoch": 0.26277372262773724,
+      "grad_norm": 9.0625,
+      "learning_rate": 2.294852310322849e-07,
+      "loss": 1.519069766998291,
+      "step": 45
+    },
+    {
+      "epoch": 0.291970802919708,
+      "grad_norm": 9.75,
+      "learning_rate": 1.1837301724142192e-07,
+      "loss": 1.5700279235839845,
+      "step": 50
+    },
+    {
+      "epoch": 0.32116788321167883,
+      "grad_norm": 7.125,
+      "learning_rate": 4.119829352859694e-08,
+      "loss": 1.6012556076049804,
+      "step": 55
+    },
+    {
+      "epoch": 0.35036496350364965,
+      "grad_norm": 8.8125,
+      "learning_rate": 3.399119873917744e-09,
+      "loss": 1.5306562423706054,
+      "step": 60
+    },
+    {
+      "epoch": 0.3562043795620438,
+      "step": 61,
+      "total_flos": 2.935955062939853e+16,
+      "train_loss": 1.5556521063945332,
+      "train_runtime": 514.4769,
+      "train_samples_per_second": 0.466,
+      "train_steps_per_second": 0.119
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 61,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": false,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.935955062939853e+16,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}
--- a/training_args.bin
+++ b/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a7cda815f0a9bcfb925251d9a4f85e9f429c06f65eb01f67ca95d6eceb98df1
+size 6033
--- a/training_loss.png
+++ b/training_loss.png