Model save

README.md

@@ -0,0 +1,58 @@
+---
+base_model: Qwen/Qwen2.5-3B-Instruct
+library_name: transformers
+model_name: Agentic-Qwen-3B-e7-lr4-b128
+tags:
+- generated_from_trainer
+- trl
+- sft
+licence: license
+---
+
+# Model Card for Agentic-Qwen-3B-e7-lr4-b128
+
+This model is a fine-tuned version of [Qwen/Qwen2.5-3B-Instruct](https://huggingface.co/Qwen/Qwen2.5-3B-Instruct).
+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="akseljoonas/Agentic-Qwen-3B-e7-lr4-b128", 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/akseljoonas-university-of-groningen/huggingface/runs/txqwvl96) 
+
+
+This model was trained with SFT.
+
+### Framework versions
+
+- TRL: 0.16.0
+- Transformers: 4.52.4
+- Pytorch: 2.6.0
+- Datasets: 3.6.0
+- 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 Gallouédec},
+	year         = 2020,
+	journal      = {GitHub repository},
+	publisher    = {GitHub},
+	howpublished = {\url{https://github.com/huggingface/trl}}
+}
+```

all_results.json

@@ -0,0 +1,8 @@
+{
+    "total_flos": 100971315986432.0,
+    "train_loss": 0.47861221517835345,
+    "train_runtime": 2363.9669,
+    "train_samples": 1845,
+    "train_samples_per_second": 5.463,
+    "train_steps_per_second": 0.044
+}

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": 1.05,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.52.4"
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "total_flos": 100971315986432.0,
+    "train_loss": 0.47861221517835345,
+    "train_runtime": 2363.9669,
+    "train_samples": 1845,
+    "train_samples_per_second": 5.463,
+    "train_steps_per_second": 0.044
+}

trainer_state.json

@@ -0,0 +1,232 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 7.0,
+  "eval_steps": 500,
+  "global_step": 105,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.3463203463203463,
+      "grad_norm": 3.5412505406821593,
+      "learning_rate": 1.4545454545454546e-05,
+      "loss": 1.6988,
+      "mean_token_accuracy": 0.6793770335614682,
+      "num_tokens": 3541090.0,
+      "step": 5
+    },
+    {
+      "epoch": 0.6926406926406926,
+      "grad_norm": 1.3307634909087853,
+      "learning_rate": 3.272727272727273e-05,
+      "loss": 0.943,
+      "mean_token_accuracy": 0.7835122548043728,
+      "num_tokens": 6936595.0,
+      "step": 10
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.6973636912379508,
+      "learning_rate": 3.989955630177008e-05,
+      "loss": 0.5573,
+      "mean_token_accuracy": 0.8710439750846003,
+      "num_tokens": 9820463.0,
+      "step": 15
+    },
+    {
+      "epoch": 1.3463203463203464,
+      "grad_norm": 0.26836269461744894,
+      "learning_rate": 3.928938350108754e-05,
+      "loss": 0.5205,
+      "mean_token_accuracy": 0.8797588124871254,
+      "num_tokens": 13203490.0,
+      "step": 20
+    },
+    {
+      "epoch": 1.6926406926406927,
+      "grad_norm": 0.20332810797148027,
+      "learning_rate": 3.814181827468682e-05,
+      "loss": 0.4801,
+      "mean_token_accuracy": 0.891349820047617,
+      "num_tokens": 16677343.0,
+      "step": 25
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 0.2671590064688769,
+      "learning_rate": 3.648883120683521e-05,
+      "loss": 0.5284,
+      "mean_token_accuracy": 0.8828430629112352,
+      "num_tokens": 19657455.0,
+      "step": 30
+    },
+    {
+      "epoch": 2.346320346320346,
+      "grad_norm": 0.20693955449412132,
+      "learning_rate": 3.437647367755859e-05,
+      "loss": 0.4191,
+      "mean_token_accuracy": 0.8993200801312924,
+      "num_tokens": 23134171.0,
+      "step": 35
+    },
+    {
+      "epoch": 2.6926406926406927,
+      "grad_norm": 0.1774648188329361,
+      "learning_rate": 3.1863594894587106e-05,
+      "loss": 0.4496,
+      "mean_token_accuracy": 0.8985331632196903,
+      "num_tokens": 26513134.0,
+      "step": 40
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 0.4177218279068929,
+      "learning_rate": 2.902020238432204e-05,
+      "loss": 0.4549,
+      "mean_token_accuracy": 0.8981883928809368,
+      "num_tokens": 29483157.0,
+      "step": 45
+    },
+    {
+      "epoch": 3.346320346320346,
+      "grad_norm": 0.21114702956816003,
+      "learning_rate": 2.5925511617712683e-05,
+      "loss": 0.3901,
+      "mean_token_accuracy": 0.9125820614397526,
+      "num_tokens": 32726053.0,
+      "step": 50
+    },
+    {
+      "epoch": 3.6926406926406927,
+      "grad_norm": 0.18992294735744386,
+      "learning_rate": 2.266573910747558e-05,
+      "loss": 0.3934,
+      "mean_token_accuracy": 0.9065288573503494,
+      "num_tokens": 36258177.0,
+      "step": 55
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 0.16760209696806955,
+      "learning_rate": 1.9331700459846516e-05,
+      "loss": 0.3554,
+      "mean_token_accuracy": 0.9131698499263172,
+      "num_tokens": 39301701.0,
+      "step": 60
+    },
+    {
+      "epoch": 4.346320346320346,
+      "grad_norm": 0.19869452464088533,
+      "learning_rate": 1.6016280297923283e-05,
+      "loss": 0.3988,
+      "mean_token_accuracy": 0.9091347955167294,
+      "num_tokens": 42943792.0,
+      "step": 65
+    },
+    {
+      "epoch": 4.692640692640692,
+      "grad_norm": 0.19890119756001573,
+      "learning_rate": 1.2811844543249748e-05,
+      "loss": 0.3305,
+      "mean_token_accuracy": 0.9202901668846607,
+      "num_tokens": 46472250.0,
+      "step": 70
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 0.3010000943274728,
+      "learning_rate": 9.80766714816165e-06,
+      "loss": 0.2949,
+      "mean_token_accuracy": 0.9255985239861717,
+      "num_tokens": 49176637.0,
+      "step": 75
+    },
+    {
+      "epoch": 5.346320346320346,
+      "grad_norm": 0.3034635853239827,
+      "learning_rate": 7.087442968823952e-06,
+      "loss": 0.3285,
+      "mean_token_accuracy": 0.9233969308435916,
+      "num_tokens": 52544587.0,
+      "step": 80
+    },
+    {
+      "epoch": 5.692640692640692,
+      "grad_norm": 0.17030309711200806,
+      "learning_rate": 4.726956069053361e-06,
+      "loss": 0.3071,
+      "mean_token_accuracy": 0.9270207457244396,
+      "num_tokens": 56073263.0,
+      "step": 85
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 0.2087495566049106,
+      "learning_rate": 2.791968414797217e-06,
+      "loss": 0.2977,
+      "mean_token_accuracy": 0.9275510193596423,
+      "num_tokens": 59025335.0,
+      "step": 90
+    },
+    {
+      "epoch": 6.346320346320346,
+      "grad_norm": 0.18514339957848594,
+      "learning_rate": 1.3363877791679491e-06,
+      "loss": 0.3248,
+      "mean_token_accuracy": 0.923822020739317,
+      "num_tokens": 62643177.0,
+      "step": 95
+    },
+    {
+      "epoch": 6.692640692640692,
+      "grad_norm": 0.1707814026740192,
+      "learning_rate": 4.0076589926826503e-07,
+      "loss": 0.2949,
+      "mean_token_accuracy": 0.9302002415060997,
+      "num_tokens": 66032974.0,
+      "step": 100
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 0.2929642273801856,
+      "learning_rate": 1.1168725394907764e-08,
+      "loss": 0.283,
+      "mean_token_accuracy": 0.9316781381486168,
+      "num_tokens": 68862598.0,
+      "step": 105
+    },
+    {
+      "epoch": 7.0,
+      "step": 105,
+      "total_flos": 100971315986432.0,
+      "train_loss": 0.47861221517835345,
+      "train_runtime": 2363.9669,
+      "train_samples_per_second": 5.463,
+      "train_steps_per_second": 0.044
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 105,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 7,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 100971315986432.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}