Model save

README.md

@@ -0,0 +1,57 @@
+---
+library_name: transformers
+model_name: Qwen-code-7B-SFT-100k-v2-cot-v5
+tags:
+- generated_from_trainer
+- trl
+- sft
+licence: license
+---
+
+# Model Card for Qwen-code-7B-SFT-100k-v2-cot-v5
+
+This model is a fine-tuned version of [None](https://huggingface.co/None).
+It has been trained using [TRL](https://github.com/huggingface/trl).
+
+## Quick start
+
+```python
+from transformers import pipeline
+
+question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
+generator = pipeline("text-generation", model="ZhuangXialie/Qwen-code-7B-SFT-100k-v2-cot-v5", device="cuda")
+output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
+print(output["generated_text"])
+```
+
+## Training procedure
+
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/dyx_team/huggingface/runs/b55zzbcb) 
+
+
+This model was trained with SFT.
+
+### Framework versions
+
+- TRL: 0.17.0.dev0
+- Transformers: 4.51.2
+- Pytorch: 2.6.0
+- Datasets: 3.5.1
+- Tokenizers: 0.21.1
+
+## Citations
+
+
+
+Cite TRL as:
+    
+```bibtex
+@misc{vonwerra2022trl,
+	title        = {{TRL: Transformer Reinforcement Learning}},
+	author       = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
+	year         = 2020,
+	journal      = {GitHub repository},
+	publisher    = {GitHub},
+	howpublished = {\url{https://github.com/huggingface/trl}}
+}
+```

all_results.json

@@ -0,0 +1,8 @@
+{
+    "total_flos": 5.6666267347582976e+17,
+    "train_loss": 0.04020713732367562,
+    "train_runtime": 749.347,
+    "train_samples": 9428,
+    "train_samples_per_second": 2.637,
+    "train_steps_per_second": 0.165
+}

generation_config.json

@@ -0,0 +1,14 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.1,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.51.2"
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "total_flos": 5.6666267347582976e+17,
+    "train_loss": 0.04020713732367562,
+    "train_runtime": 749.347,
+    "train_samples": 9428,
+    "train_samples_per_second": 2.637,
+    "train_steps_per_second": 0.165
+}

trainer_state.json

@@ -0,0 +1,261 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 4.0,
+  "eval_steps": 500,
+  "global_step": 124,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.16129032258064516,
+      "grad_norm": 1.9339179610924326,
+      "learning_rate": 1.5384615384615387e-05,
+      "loss": 0.2023,
+      "mean_token_accuracy": 0.9648663103580475,
+      "num_tokens": 1200802.0,
+      "step": 5
+    },
+    {
+      "epoch": 0.3225806451612903,
+      "grad_norm": 0.20186157671436705,
+      "learning_rate": 3.461538461538462e-05,
+      "loss": 0.0907,
+      "mean_token_accuracy": 0.979491549730301,
+      "num_tokens": 2329939.0,
+      "step": 10
+    },
+    {
+      "epoch": 0.4838709677419355,
+      "grad_norm": 0.11105863000001426,
+      "learning_rate": 4.999098891016224e-05,
+      "loss": 0.062,
+      "mean_token_accuracy": 0.9859958469867707,
+      "num_tokens": 3454242.0,
+      "step": 15
+    },
+    {
+      "epoch": 0.6451612903225806,
+      "grad_norm": 0.07295180708170615,
+      "learning_rate": 4.967635798282344e-05,
+      "loss": 0.058,
+      "mean_token_accuracy": 0.9865534484386445,
+      "num_tokens": 4636563.0,
+      "step": 20
+    },
+    {
+      "epoch": 0.8064516129032258,
+      "grad_norm": 0.061430086599508354,
+      "learning_rate": 4.8918364374405345e-05,
+      "loss": 0.0517,
+      "mean_token_accuracy": 0.9883540272712708,
+      "num_tokens": 5780954.0,
+      "step": 25
+    },
+    {
+      "epoch": 0.967741935483871,
+      "grad_norm": 0.06636359184947321,
+      "learning_rate": 4.773216233515633e-05,
+      "loss": 0.0523,
+      "mean_token_accuracy": 0.9879270672798157,
+      "num_tokens": 6973730.0,
+      "step": 30
+    },
+    {
+      "epoch": 1.129032258064516,
+      "grad_norm": 0.05553079305562554,
+      "learning_rate": 4.614146710798645e-05,
+      "loss": 0.0441,
+      "mean_token_accuracy": 0.9883361339569092,
+      "num_tokens": 8106792.0,
+      "step": 35
+    },
+    {
+      "epoch": 1.2903225806451613,
+      "grad_norm": 0.05845614425520945,
+      "learning_rate": 4.417808079950151e-05,
+      "loss": 0.0444,
+      "mean_token_accuracy": 0.9887041985988617,
+      "num_tokens": 9259980.0,
+      "step": 40
+    },
+    {
+      "epoch": 1.4516129032258065,
+      "grad_norm": 0.05205247747909711,
+      "learning_rate": 4.188125657374216e-05,
+      "loss": 0.0389,
+      "mean_token_accuracy": 0.990032184123993,
+      "num_tokens": 10469264.0,
+      "step": 45
+    },
+    {
+      "epoch": 1.6129032258064515,
+      "grad_norm": 0.06709443615050471,
+      "learning_rate": 3.929691388003772e-05,
+      "loss": 0.0419,
+      "mean_token_accuracy": 0.989645528793335,
+      "num_tokens": 11528973.0,
+      "step": 50
+    },
+    {
+      "epoch": 1.7741935483870968,
+      "grad_norm": 0.049780936243340075,
+      "learning_rate": 3.647672040459687e-05,
+      "loss": 0.0308,
+      "mean_token_accuracy": 0.9919266939163208,
+      "num_tokens": 12714383.0,
+      "step": 55
+    },
+    {
+      "epoch": 1.935483870967742,
+      "grad_norm": 0.05777914245949305,
+      "learning_rate": 3.347705909999472e-05,
+      "loss": 0.0345,
+      "mean_token_accuracy": 0.9910649061203003,
+      "num_tokens": 13828224.0,
+      "step": 60
+    },
+    {
+      "epoch": 2.096774193548387,
+      "grad_norm": 0.0763532245250884,
+      "learning_rate": 3.0357900944304774e-05,
+      "loss": 0.0298,
+      "mean_token_accuracy": 0.9922283053398132,
+      "num_tokens": 14895428.0,
+      "step": 65
+    },
+    {
+      "epoch": 2.258064516129032,
+      "grad_norm": 0.0478737582615227,
+      "learning_rate": 2.7181605966332857e-05,
+      "loss": 0.0249,
+      "mean_token_accuracy": 0.9926450073719024,
+      "num_tokens": 16001285.0,
+      "step": 70
+    },
+    {
+      "epoch": 2.4193548387096775,
+      "grad_norm": 0.059113621264175666,
+      "learning_rate": 2.4011676507555546e-05,
+      "loss": 0.0231,
+      "mean_token_accuracy": 0.9934819102287292,
+      "num_tokens": 17101738.0,
+      "step": 75
+    },
+    {
+      "epoch": 2.5806451612903225,
+      "grad_norm": 0.0666372605177016,
+      "learning_rate": 2.0911487646277623e-05,
+      "loss": 0.0232,
+      "mean_token_accuracy": 0.9933121263980865,
+      "num_tokens": 18330596.0,
+      "step": 80
+    },
+    {
+      "epoch": 2.741935483870968,
+      "grad_norm": 0.06526290734512033,
+      "learning_rate": 1.7943020166108926e-05,
+      "loss": 0.0252,
+      "mean_token_accuracy": 0.9927115201950073,
+      "num_tokens": 19460189.0,
+      "step": 85
+    },
+    {
+      "epoch": 2.903225806451613,
+      "grad_norm": 0.06160045585242219,
+      "learning_rate": 1.5165621399994035e-05,
+      "loss": 0.0227,
+      "mean_token_accuracy": 0.9932076632976532,
+      "num_tokens": 20686498.0,
+      "step": 90
+    },
+    {
+      "epoch": 3.064516129032258,
+      "grad_norm": 0.05412733531697612,
+      "learning_rate": 1.2634818723723174e-05,
+      "loss": 0.0215,
+      "mean_token_accuracy": 0.9935298144817353,
+      "num_tokens": 21800001.0,
+      "step": 95
+    },
+    {
+      "epoch": 3.225806451612903,
+      "grad_norm": 0.057286274686510985,
+      "learning_rate": 1.0401209420254312e-05,
+      "loss": 0.0123,
+      "mean_token_accuracy": 0.9964357137680053,
+      "num_tokens": 23023063.0,
+      "step": 100
+    },
+    {
+      "epoch": 3.3870967741935485,
+      "grad_norm": 0.06710958627001332,
+      "learning_rate": 8.509449109326118e-06,
+      "loss": 0.0121,
+      "mean_token_accuracy": 0.9962392687797547,
+      "num_tokens": 24175633.0,
+      "step": 105
+    },
+    {
+      "epoch": 3.5483870967741935,
+      "grad_norm": 0.051561033179333415,
+      "learning_rate": 6.9973589662669455e-06,
+      "loss": 0.0129,
+      "mean_token_accuracy": 0.9961234986782074,
+      "num_tokens": 25391505.0,
+      "step": 110
+    },
+    {
+      "epoch": 3.709677419354839,
+      "grad_norm": 0.07447115032041235,
+      "learning_rate": 5.895169579001987e-06,
+      "loss": 0.0137,
+      "mean_token_accuracy": 0.9959305047988891,
+      "num_tokens": 26515619.0,
+      "step": 115
+    },
+    {
+      "epoch": 3.870967741935484,
+      "grad_norm": 0.05952950063195277,
+      "learning_rate": 5.224916560510316e-06,
+      "loss": 0.0133,
+      "mean_token_accuracy": 0.9959637641906738,
+      "num_tokens": 27647513.0,
+      "step": 120
+    },
+    {
+      "epoch": 4.0,
+      "mean_token_accuracy": 0.996899887919426,
+      "num_tokens": 28505880.0,
+      "step": 124,
+      "total_flos": 5.6666267347582976e+17,
+      "train_loss": 0.04020713732367562,
+      "train_runtime": 749.347,
+      "train_samples_per_second": 2.637,
+      "train_steps_per_second": 0.165
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 124,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 1000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 5.6666267347582976e+17,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}