Upload BF-Router v0.5 QLoRA adapter (Qwen3-4B)

.gitattributes

@@ -33,3 +33,7 @@ saved_model/**/* 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
+checkpoint-281/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-562/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-843/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,62 @@
+---
+base_model: Qwen/Qwen3-4B-Instruct-2507
+library_name: peft
+model_name: bf-router-v0.5
+tags:
+- base_model:adapter:Qwen/Qwen3-4B-Instruct-2507
+- lora
+- sft
+- transformers
+- trl
+licence: license
+pipeline_tag: text-generation
+---
+
+# Model Card for bf-router-v0.5
+
+This model is a fine-tuned version of [Qwen/Qwen3-4B-Instruct-2507](https://huggingface.co/Qwen/Qwen3-4B-Instruct-2507).
+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="None", device="cuda")
+output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
+print(output["generated_text"])
+```
+
+## Training procedure
+
+ 
+
+
+
+This model was trained with SFT.
+
+### Framework versions
+
+- PEFT 0.18.1
+- TRL: 0.29.0
+- Transformers: 5.2.0
+- Pytorch: 2.10.0
+- Datasets: 4.6.1
+- Tokenizers: 0.22.2
+
+## Citations
+
+
+
+Cite TRL as:
+    
+```bibtex
+@software{vonwerra2020trl,
+  title   = {{TRL: Transformers Reinforcement Learning}},
+  author  = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin},
+  license = {Apache-2.0},
+  url     = {https://github.com/huggingface/trl},
+  year    = {2020}
+}
+```

adapter_config.json

@@ -0,0 +1,46 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3-4B-Instruct-2507",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.18.1",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "down_proj",
+    "up_proj",
+    "q_proj",
+    "gate_proj",
+    "k_proj",
+    "o_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0bc1f22f429ab6ab7527f80ca2f343b90e4ea80c39dda2ec83bada6e7924d676
+size 132187888

chat_template.jinja

@@ -0,0 +1,61 @@
+{%- 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 %}
+{%- 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" %}
+        {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- 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' }}
+{%- endif %}

checkpoint-281/README.md

@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-4B-Instruct-2507
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-4B-Instruct-2507
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1

checkpoint-281/adapter_config.json

@@ -0,0 +1,46 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3-4B-Instruct-2507",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.18.1",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "down_proj",
+    "up_proj",
+    "q_proj",
+    "gate_proj",
+    "k_proj",
+    "o_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

checkpoint-281/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:581d2f686c971b84c9ab741950e24777fcd2dc0452a11ae7df7b862883fc2804
+size 66127776

checkpoint-281/chat_template.jinja

@@ -0,0 +1,61 @@
+{%- 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 %}
+{%- 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" %}
+        {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- 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' }}
+{%- endif %}

checkpoint-281/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:875ffb16cc18b423e5ef0a71bd54a2a49974c68c9ee717b877987df1945c5e51
+size 132552715

checkpoint-281/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:95e70b1e3fd60fe41298a7546c09ee7f379684cfb911eba36e8ca1865e2f5d46
+size 14645

checkpoint-281/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aeccf90c5a1ed591a7458cc9cf95f5fb44b5b6e02069bc531935e2d480503332
+size 1465

checkpoint-281/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
+size 11422650

checkpoint-281/tokenizer_config.json

@@ -0,0 +1,29 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "model_max_length": 1010000,
+  "pad_token": "<|im_end|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoint-281/trainer_state.json

@@ -0,0 +1,335 @@
+{
+  "best_global_step": 281,
+  "best_metric": 0.18149949610233307,
+  "best_model_checkpoint": "/app/output/bf-router-v0.5/checkpoint-281",
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 281,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.4157426357269287,
+      "epoch": 0.003561887800534283,
+      "grad_norm": 1.203125,
+      "learning_rate": 0.0,
+      "loss": 4.336674213409424,
+      "mean_token_accuracy": 0.4680640399456024,
+      "num_tokens": 3092.0,
+      "step": 1
+    },
+    {
+      "entropy": 1.5884604156017303,
+      "epoch": 0.03561887800534283,
+      "grad_norm": 0.59375,
+      "learning_rate": 6.923076923076924e-05,
+      "loss": 3.986017862955729,
+      "mean_token_accuracy": 0.4735516524977154,
+      "num_tokens": 30650.0,
+      "step": 10
+    },
+    {
+      "entropy": 2.2315500736236573,
+      "epoch": 0.07123775601068566,
+      "grad_norm": 0.26953125,
+      "learning_rate": 0.00014615384615384615,
+      "loss": 2.16544189453125,
+      "mean_token_accuracy": 0.6272616922855377,
+      "num_tokens": 61735.0,
+      "step": 20
+    },
+    {
+      "entropy": 0.8066075079143047,
+      "epoch": 0.10685663401602849,
+      "grad_norm": 0.298828125,
+      "learning_rate": 0.00019999334629824895,
+      "loss": 0.7935717582702637,
+      "mean_token_accuracy": 0.857068446278572,
+      "num_tokens": 92484.0,
+      "step": 30
+    },
+    {
+      "entropy": 0.5340980306267739,
+      "epoch": 0.14247551202137132,
+      "grad_norm": 0.10693359375,
+      "learning_rate": 0.00019987508289735716,
+      "loss": 0.5261529922485352,
+      "mean_token_accuracy": 0.9004581302404404,
+      "num_tokens": 123010.0,
+      "step": 40
+    },
+    {
+      "entropy": 0.4001643668860197,
+      "epoch": 0.17809439002671415,
+      "grad_norm": 0.08251953125,
+      "learning_rate": 0.0001996091607179287,
+      "loss": 0.3956298351287842,
+      "mean_token_accuracy": 0.9209299519658088,
+      "num_tokens": 153710.0,
+      "step": 50
+    },
+    {
+      "entropy": 0.34828933551907537,
+      "epoch": 0.21371326803205698,
+      "grad_norm": 0.10009765625,
+      "learning_rate": 0.00019919597290851538,
+      "loss": 0.3648662090301514,
+      "mean_token_accuracy": 0.9263768374919892,
+      "num_tokens": 184577.0,
+      "step": 60
+    },
+    {
+      "entropy": 0.2983129292726517,
+      "epoch": 0.2493321460373998,
+      "grad_norm": 0.08447265625,
+      "learning_rate": 0.00019863613034027224,
+      "loss": 0.29642860889434813,
+      "mean_token_accuracy": 0.9371836110949516,
+      "num_tokens": 215040.0,
+      "step": 70
+    },
+    {
+      "entropy": 0.280751570686698,
+      "epoch": 0.28495102404274264,
+      "grad_norm": 0.099609375,
+      "learning_rate": 0.00019793046070382437,
+      "loss": 0.2813178300857544,
+      "mean_token_accuracy": 0.9388988897204399,
+      "num_tokens": 245683.0,
+      "step": 80
+    },
+    {
+      "entropy": 0.2960826952010393,
+      "epoch": 0.3205699020480855,
+      "grad_norm": 0.11962890625,
+      "learning_rate": 0.00019708000728558064,
+      "loss": 0.31115410327911375,
+      "mean_token_accuracy": 0.9335825845599175,
+      "num_tokens": 276430.0,
+      "step": 90
+    },
+    {
+      "entropy": 0.28416929207742214,
+      "epoch": 0.3561887800534283,
+      "grad_norm": 0.103515625,
+      "learning_rate": 0.00019608602742530283,
+      "loss": 0.27938365936279297,
+      "mean_token_accuracy": 0.9393619552254677,
+      "num_tokens": 307296.0,
+      "step": 100
+    },
+    {
+      "entropy": 0.23417662251740695,
+      "epoch": 0.39180765805877116,
+      "grad_norm": 0.06298828125,
+      "learning_rate": 0.00019494999065721108,
+      "loss": 0.23734774589538574,
+      "mean_token_accuracy": 0.9487677246332169,
+      "num_tokens": 337962.0,
+      "step": 110
+    },
+    {
+      "entropy": 0.24019648060202597,
+      "epoch": 0.42742653606411396,
+      "grad_norm": 0.06640625,
+      "learning_rate": 0.0001936735765373737,
+      "loss": 0.245324969291687,
+      "mean_token_accuracy": 0.9457737937569618,
+      "num_tokens": 368782.0,
+      "step": 120
+    },
+    {
+      "entropy": 0.269666058011353,
+      "epoch": 0.4630454140694568,
+      "grad_norm": 0.09765625,
+      "learning_rate": 0.00019225867216059325,
+      "loss": 0.2686375617980957,
+      "mean_token_accuracy": 0.9398508563637733,
+      "num_tokens": 399329.0,
+      "step": 130
+    },
+    {
+      "entropy": 0.18328354582190515,
+      "epoch": 0.4986642920747996,
+      "grad_norm": 0.06298828125,
+      "learning_rate": 0.00019070736937046035,
+      "loss": 0.18070420026779174,
+      "mean_token_accuracy": 0.9568557634949684,
+      "num_tokens": 429759.0,
+      "step": 140
+    },
+    {
+      "entropy": 0.21165105439722537,
+      "epoch": 0.5342831700801425,
+      "grad_norm": 0.0810546875,
+      "learning_rate": 0.0001890219616666997,
+      "loss": 0.22004930973052977,
+      "mean_token_accuracy": 0.9494227185845375,
+      "num_tokens": 460409.0,
+      "step": 150
+    },
+    {
+      "entropy": 0.23916025627404452,
+      "epoch": 0.5699020480854853,
+      "grad_norm": 0.07568359375,
+      "learning_rate": 0.00018720494081438078,
+      "loss": 0.2379377841949463,
+      "mean_token_accuracy": 0.9455192387104034,
+      "num_tokens": 491344.0,
+      "step": 160
+    },
+    {
+      "entropy": 0.2238690486177802,
+      "epoch": 0.6055209260908282,
+      "grad_norm": 0.08984375,
+      "learning_rate": 0.00018525899316000608,
+      "loss": 0.2245168685913086,
+      "mean_token_accuracy": 0.9485938593745231,
+      "num_tokens": 522029.0,
+      "step": 170
+    },
+    {
+      "entropy": 0.17032606173306702,
+      "epoch": 0.641139804096171,
+      "grad_norm": 0.080078125,
+      "learning_rate": 0.00018318699565992357,
+      "loss": 0.17549347877502441,
+      "mean_token_accuracy": 0.9574712902307511,
+      "num_tokens": 552306.0,
+      "step": 180
+    },
+    {
+      "entropy": 0.21890966054052113,
+      "epoch": 0.6767586821015138,
+      "grad_norm": 0.0673828125,
+      "learning_rate": 0.00018099201162693476,
+      "loss": 0.21968293190002441,
+      "mean_token_accuracy": 0.9482985377311707,
+      "num_tokens": 583181.0,
+      "step": 190
+    },
+    {
+      "entropy": 0.20646683853119613,
+      "epoch": 0.7123775601068566,
+      "grad_norm": 0.09765625,
+      "learning_rate": 0.00017867728620138708,
+      "loss": 0.20968315601348878,
+      "mean_token_accuracy": 0.9519409075379371,
+      "num_tokens": 613818.0,
+      "step": 200
+    },
+    {
+      "entropy": 0.23000301327556372,
+      "epoch": 0.7479964381121995,
+      "grad_norm": 0.08935546875,
+      "learning_rate": 0.00017624624155344626,
+      "loss": 0.2383474349975586,
+      "mean_token_accuracy": 0.9458372846245766,
+      "num_tokens": 644576.0,
+      "step": 210
+    },
+    {
+      "entropy": 0.21420854832977057,
+      "epoch": 0.7836153161175423,
+      "grad_norm": 0.0966796875,
+      "learning_rate": 0.0001737024718236413,
+      "loss": 0.21017465591430665,
+      "mean_token_accuracy": 0.9511569887399673,
+      "num_tokens": 675564.0,
+      "step": 220
+    },
+    {
+      "entropy": 0.19280126914381981,
+      "epoch": 0.8192341941228851,
+      "grad_norm": 0.08154296875,
+      "learning_rate": 0.00017104973780916294,
+      "loss": 0.1936139941215515,
+      "mean_token_accuracy": 0.954449312388897,
+      "num_tokens": 706309.0,
+      "step": 230
+    },
+    {
+      "entropy": 0.22524890769273043,
+      "epoch": 0.8548530721282279,
+      "grad_norm": 0.10546875,
+      "learning_rate": 0.00016829196140377085,
+      "loss": 0.22950620651245118,
+      "mean_token_accuracy": 0.9476036429405212,
+      "num_tokens": 737135.0,
+      "step": 240
+    },
+    {
+      "entropy": 0.18842433094978334,
+      "epoch": 0.8904719501335708,
+      "grad_norm": 0.087890625,
+      "learning_rate": 0.00016543321979953007,
+      "loss": 0.19049547910690307,
+      "mean_token_accuracy": 0.9543813273310662,
+      "num_tokens": 767853.0,
+      "step": 250
+    },
+    {
+      "entropy": 0.1991994746029377,
+      "epoch": 0.9260908281389136,
+      "grad_norm": 0.07763671875,
+      "learning_rate": 0.00016247773945894962,
+      "loss": 0.19527161121368408,
+      "mean_token_accuracy": 0.9535173490643502,
+      "num_tokens": 798511.0,
+      "step": 260
+    },
+    {
+      "entropy": 0.18196075949817897,
+      "epoch": 0.9617097061442564,
+      "grad_norm": 0.0771484375,
+      "learning_rate": 0.00015942988986643352,
+      "loss": 0.19091761112213135,
+      "mean_token_accuracy": 0.9535432115197182,
+      "num_tokens": 829447.0,
+      "step": 270
+    },
+    {
+      "entropy": 0.18307169582694768,
+      "epoch": 0.9973285841495992,
+      "grad_norm": 0.08837890625,
+      "learning_rate": 0.00015629417706828423,
+      "loss": 0.1818631410598755,
+      "mean_token_accuracy": 0.955574706196785,
+      "num_tokens": 860142.0,
+      "step": 280
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.18207345770817276,
+      "eval_loss": 0.18149949610233307,
+      "eval_mean_token_accuracy": 0.9555717660180221,
+      "eval_num_tokens": 862512.0,
+      "eval_runtime": 35.3256,
+      "eval_samples_per_second": 15.881,
+      "eval_steps_per_second": 3.991,
+      "step": 281
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 843,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.028075887327232e+16,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-281/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:57851b427cb3a4649963b62f1b5d8860aa50ba7ac4abbbb2ebc56acbf65bd694
+size 5649

checkpoint-562/README.md

@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-4B-Instruct-2507
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-4B-Instruct-2507
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1

checkpoint-562/adapter_config.json

@@ -0,0 +1,46 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3-4B-Instruct-2507",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.18.1",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "down_proj",
+    "up_proj",
+    "q_proj",
+    "gate_proj",
+    "k_proj",
+    "o_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

checkpoint-562/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ae3fd5d71a6d75e54a11294dc7fb4c6024c2786654ab088f72658e12516606c7
+size 66127776

checkpoint-562/chat_template.jinja

@@ -0,0 +1,61 @@
+{%- 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 %}
+{%- 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" %}
+        {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- 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' }}
+{%- endif %}

checkpoint-562/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9367da0e9e0f21638be2669e6d4ae2225b3503820ad3583a3c9b0334bf37c92d
+size 132552715

checkpoint-562/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4313f9d832b6c6d568b3ac0b53ef1b8e83d1072bbf993a8167c50a30e46375a7
+size 14645

checkpoint-562/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d813cca777bf0ff28660547edc526875199f3774d19049acf9916e60c27f3d6
+size 1465

checkpoint-562/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
+size 11422650

checkpoint-562/tokenizer_config.json

@@ -0,0 +1,29 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "model_max_length": 1010000,
+  "pad_token": "<|im_end|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoint-562/trainer_state.json

@@ -0,0 +1,626 @@
+{
+  "best_global_step": 562,
+  "best_metric": 0.15477769076824188,
+  "best_model_checkpoint": "/app/output/bf-router-v0.5/checkpoint-562",
+  "epoch": 2.0,
+  "eval_steps": 500,
+  "global_step": 562,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.4157426357269287,
+      "epoch": 0.003561887800534283,
+      "grad_norm": 1.203125,
+      "learning_rate": 0.0,
+      "loss": 4.336674213409424,
+      "mean_token_accuracy": 0.4680640399456024,
+      "num_tokens": 3092.0,
+      "step": 1
+    },
+    {
+      "entropy": 1.5884604156017303,
+      "epoch": 0.03561887800534283,
+      "grad_norm": 0.59375,
+      "learning_rate": 6.923076923076924e-05,
+      "loss": 3.986017862955729,
+      "mean_token_accuracy": 0.4735516524977154,
+      "num_tokens": 30650.0,
+      "step": 10
+    },
+    {
+      "entropy": 2.2315500736236573,
+      "epoch": 0.07123775601068566,
+      "grad_norm": 0.26953125,
+      "learning_rate": 0.00014615384615384615,
+      "loss": 2.16544189453125,
+      "mean_token_accuracy": 0.6272616922855377,
+      "num_tokens": 61735.0,
+      "step": 20
+    },
+    {
+      "entropy": 0.8066075079143047,
+      "epoch": 0.10685663401602849,
+      "grad_norm": 0.298828125,
+      "learning_rate": 0.00019999334629824895,
+      "loss": 0.7935717582702637,
+      "mean_token_accuracy": 0.857068446278572,
+      "num_tokens": 92484.0,
+      "step": 30
+    },
+    {
+      "entropy": 0.5340980306267739,
+      "epoch": 0.14247551202137132,
+      "grad_norm": 0.10693359375,
+      "learning_rate": 0.00019987508289735716,
+      "loss": 0.5261529922485352,
+      "mean_token_accuracy": 0.9004581302404404,
+      "num_tokens": 123010.0,
+      "step": 40
+    },
+    {
+      "entropy": 0.4001643668860197,
+      "epoch": 0.17809439002671415,
+      "grad_norm": 0.08251953125,
+      "learning_rate": 0.0001996091607179287,
+      "loss": 0.3956298351287842,
+      "mean_token_accuracy": 0.9209299519658088,
+      "num_tokens": 153710.0,
+      "step": 50
+    },
+    {
+      "entropy": 0.34828933551907537,
+      "epoch": 0.21371326803205698,
+      "grad_norm": 0.10009765625,
+      "learning_rate": 0.00019919597290851538,
+      "loss": 0.3648662090301514,
+      "mean_token_accuracy": 0.9263768374919892,
+      "num_tokens": 184577.0,
+      "step": 60
+    },
+    {
+      "entropy": 0.2983129292726517,
+      "epoch": 0.2493321460373998,
+      "grad_norm": 0.08447265625,
+      "learning_rate": 0.00019863613034027224,
+      "loss": 0.29642860889434813,
+      "mean_token_accuracy": 0.9371836110949516,
+      "num_tokens": 215040.0,
+      "step": 70
+    },
+    {
+      "entropy": 0.280751570686698,
+      "epoch": 0.28495102404274264,
+      "grad_norm": 0.099609375,
+      "learning_rate": 0.00019793046070382437,
+      "loss": 0.2813178300857544,
+      "mean_token_accuracy": 0.9388988897204399,
+      "num_tokens": 245683.0,
+      "step": 80
+    },
+    {
+      "entropy": 0.2960826952010393,
+      "epoch": 0.3205699020480855,
+      "grad_norm": 0.11962890625,
+      "learning_rate": 0.00019708000728558064,
+      "loss": 0.31115410327911375,
+      "mean_token_accuracy": 0.9335825845599175,
+      "num_tokens": 276430.0,
+      "step": 90
+    },
+    {
+      "entropy": 0.28416929207742214,
+      "epoch": 0.3561887800534283,
+      "grad_norm": 0.103515625,
+      "learning_rate": 0.00019608602742530283,
+      "loss": 0.27938365936279297,
+      "mean_token_accuracy": 0.9393619552254677,
+      "num_tokens": 307296.0,
+      "step": 100
+    },
+    {
+      "entropy": 0.23417662251740695,
+      "epoch": 0.39180765805877116,
+      "grad_norm": 0.06298828125,
+      "learning_rate": 0.00019494999065721108,
+      "loss": 0.23734774589538574,
+      "mean_token_accuracy": 0.9487677246332169,
+      "num_tokens": 337962.0,
+      "step": 110
+    },
+    {
+      "entropy": 0.24019648060202597,
+      "epoch": 0.42742653606411396,
+      "grad_norm": 0.06640625,
+      "learning_rate": 0.0001936735765373737,
+      "loss": 0.245324969291687,
+      "mean_token_accuracy": 0.9457737937569618,
+      "num_tokens": 368782.0,
+      "step": 120
+    },
+    {
+      "entropy": 0.269666058011353,
+      "epoch": 0.4630454140694568,
+      "grad_norm": 0.09765625,
+      "learning_rate": 0.00019225867216059325,
+      "loss": 0.2686375617980957,
+      "mean_token_accuracy": 0.9398508563637733,
+      "num_tokens": 399329.0,
+      "step": 130
+    },
+    {
+      "entropy": 0.18328354582190515,
+      "epoch": 0.4986642920747996,
+      "grad_norm": 0.06298828125,
+      "learning_rate": 0.00019070736937046035,
+      "loss": 0.18070420026779174,
+      "mean_token_accuracy": 0.9568557634949684,
+      "num_tokens": 429759.0,
+      "step": 140
+    },
+    {
+      "entropy": 0.21165105439722537,
+      "epoch": 0.5342831700801425,
+      "grad_norm": 0.0810546875,
+      "learning_rate": 0.0001890219616666997,
+      "loss": 0.22004930973052977,
+      "mean_token_accuracy": 0.9494227185845375,
+      "num_tokens": 460409.0,
+      "step": 150
+    },
+    {
+      "entropy": 0.23916025627404452,
+      "epoch": 0.5699020480854853,
+      "grad_norm": 0.07568359375,
+      "learning_rate": 0.00018720494081438078,
+      "loss": 0.2379377841949463,
+      "mean_token_accuracy": 0.9455192387104034,
+      "num_tokens": 491344.0,
+      "step": 160
+    },
+    {
+      "entropy": 0.2238690486177802,
+      "epoch": 0.6055209260908282,
+      "grad_norm": 0.08984375,
+      "learning_rate": 0.00018525899316000608,
+      "loss": 0.2245168685913086,
+      "mean_token_accuracy": 0.9485938593745231,
+      "num_tokens": 522029.0,
+      "step": 170
+    },
+    {
+      "entropy": 0.17032606173306702,
+      "epoch": 0.641139804096171,
+      "grad_norm": 0.080078125,
+      "learning_rate": 0.00018318699565992357,
+      "loss": 0.17549347877502441,
+      "mean_token_accuracy": 0.9574712902307511,
+      "num_tokens": 552306.0,
+      "step": 180
+    },
+    {
+      "entropy": 0.21890966054052113,
+      "epoch": 0.6767586821015138,
+      "grad_norm": 0.0673828125,
+      "learning_rate": 0.00018099201162693476,
+      "loss": 0.21968293190002441,
+      "mean_token_accuracy": 0.9482985377311707,
+      "num_tokens": 583181.0,
+      "step": 190
+    },
+    {
+      "entropy": 0.20646683853119613,
+      "epoch": 0.7123775601068566,
+      "grad_norm": 0.09765625,
+      "learning_rate": 0.00017867728620138708,
+      "loss": 0.20968315601348878,
+      "mean_token_accuracy": 0.9519409075379371,
+      "num_tokens": 613818.0,
+      "step": 200
+    },
+    {
+      "entropy": 0.23000301327556372,
+      "epoch": 0.7479964381121995,
+      "grad_norm": 0.08935546875,
+      "learning_rate": 0.00017624624155344626,
+      "loss": 0.2383474349975586,
+      "mean_token_accuracy": 0.9458372846245766,
+      "num_tokens": 644576.0,
+      "step": 210
+    },
+    {
+      "entropy": 0.21420854832977057,
+      "epoch": 0.7836153161175423,
+      "grad_norm": 0.0966796875,
+      "learning_rate": 0.0001737024718236413,
+      "loss": 0.21017465591430665,
+      "mean_token_accuracy": 0.9511569887399673,
+      "num_tokens": 675564.0,
+      "step": 220
+    },
+    {
+      "entropy": 0.19280126914381981,
+      "epoch": 0.8192341941228851,
+      "grad_norm": 0.08154296875,
+      "learning_rate": 0.00017104973780916294,
+      "loss": 0.1936139941215515,
+      "mean_token_accuracy": 0.954449312388897,
+      "num_tokens": 706309.0,
+      "step": 230
+    },
+    {
+      "entropy": 0.22524890769273043,
+      "epoch": 0.8548530721282279,
+      "grad_norm": 0.10546875,
+      "learning_rate": 0.00016829196140377085,
+      "loss": 0.22950620651245118,
+      "mean_token_accuracy": 0.9476036429405212,
+      "num_tokens": 737135.0,
+      "step": 240
+    },
+    {
+      "entropy": 0.18842433094978334,
+      "epoch": 0.8904719501335708,
+      "grad_norm": 0.087890625,
+      "learning_rate": 0.00016543321979953007,
+      "loss": 0.19049547910690307,
+      "mean_token_accuracy": 0.9543813273310662,
+      "num_tokens": 767853.0,
+      "step": 250
+    },
+    {
+      "entropy": 0.1991994746029377,
+      "epoch": 0.9260908281389136,
+      "grad_norm": 0.07763671875,
+      "learning_rate": 0.00016247773945894962,
+      "loss": 0.19527161121368408,
+      "mean_token_accuracy": 0.9535173490643502,
+      "num_tokens": 798511.0,
+      "step": 260
+    },
+    {
+      "entropy": 0.18196075949817897,
+      "epoch": 0.9617097061442564,
+      "grad_norm": 0.0771484375,
+      "learning_rate": 0.00015942988986643352,
+      "loss": 0.19091761112213135,
+      "mean_token_accuracy": 0.9535432115197182,
+      "num_tokens": 829447.0,
+      "step": 270
+    },
+    {
+      "entropy": 0.18307169582694768,
+      "epoch": 0.9973285841495992,
+      "grad_norm": 0.08837890625,
+      "learning_rate": 0.00015629417706828423,
+      "loss": 0.1818631410598755,
+      "mean_token_accuracy": 0.955574706196785,
+      "num_tokens": 860142.0,
+      "step": 280
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.18207345770817276,
+      "eval_loss": 0.18149949610233307,
+      "eval_mean_token_accuracy": 0.9555717660180221,
+      "eval_num_tokens": 862512.0,
+      "eval_runtime": 35.3256,
+      "eval_samples_per_second": 15.881,
+      "eval_steps_per_second": 3.991,
+      "step": 281
+    },
+    {
+      "entropy": 0.1545313375118451,
+      "epoch": 1.0320569902048085,
+      "grad_norm": 0.064453125,
+      "learning_rate": 0.00015307523701080768,
+      "loss": 0.14675980806350708,
+      "mean_token_accuracy": 0.9623546004295349,
+      "num_tokens": 889907.0,
+      "step": 290
+    },
+    {
+      "entropy": 0.13382284864783286,
+      "epoch": 1.0676758682101515,
+      "grad_norm": 0.060546875,
+      "learning_rate": 0.00014977782868636999,
+      "loss": 0.13706474304199218,
+      "mean_token_accuracy": 0.9633408606052398,
+      "num_tokens": 920639.0,
+      "step": 300
+    },
+    {
+      "entropy": 0.16596811451017857,
+      "epoch": 1.1032947462154943,
+      "grad_norm": 0.0791015625,
+      "learning_rate": 0.00014640682709753832,
+      "loss": 0.16318607330322266,
+      "mean_token_accuracy": 0.9578790530562401,
+      "num_tokens": 951323.0,
+      "step": 310
+    },
+    {
+      "entropy": 0.1480622159317136,
+      "epoch": 1.138913624220837,
+      "grad_norm": 0.057861328125,
+      "learning_rate": 0.0001429672160497085,
+      "loss": 0.1488279938697815,
+      "mean_token_accuracy": 0.9609164595603943,
+      "num_tokens": 981911.0,
+      "step": 320
+    },
+    {
+      "entropy": 0.1655621325597167,
+      "epoch": 1.1745325022261799,
+      "grad_norm": 0.07275390625,
+      "learning_rate": 0.00013946408078287462,
+      "loss": 0.16427644491195678,
+      "mean_token_accuracy": 0.9576473370194435,
+      "num_tokens": 1012535.0,
+      "step": 330
+    },
+    {
+      "entropy": 0.18122829273343086,
+      "epoch": 1.2101513802315227,
+      "grad_norm": 0.06689453125,
+      "learning_rate": 0.00013590260045343432,
+      "loss": 0.176006543636322,
+      "mean_token_accuracy": 0.9551453411579132,
+      "num_tokens": 1043503.0,
+      "step": 340
+    },
+    {
+      "entropy": 0.15644419118762015,
+      "epoch": 1.2457702582368655,
+      "grad_norm": 0.06201171875,
+      "learning_rate": 0.00013228804047714463,
+      "loss": 0.1652477502822876,
+      "mean_token_accuracy": 0.9575047269463539,
+      "num_tokens": 1074168.0,
+      "step": 350
+    },
+    {
+      "entropy": 0.14828295167535543,
+      "epoch": 1.2813891362422083,
+      "grad_norm": 0.06982421875,
+      "learning_rate": 0.00012862574474454928,
+      "loss": 0.14066768884658815,
+      "mean_token_accuracy": 0.9639188721776009,
+      "num_tokens": 1104681.0,
+      "step": 360
+    },
+    {
+      "entropy": 0.19175196047872306,
+      "epoch": 1.317008014247551,
+      "grad_norm": 0.09619140625,
+      "learning_rate": 0.0001249211277203859,
+      "loss": 0.1917089343070984,
+      "mean_token_accuracy": 0.9522068575024605,
+      "num_tokens": 1135361.0,
+      "step": 370
+    },
+    {
+      "entropy": 0.1644950734451413,
+      "epoch": 1.3526268922528941,
+      "grad_norm": 0.0849609375,
+      "learning_rate": 0.00012117966643865398,
+      "loss": 0.16864393949508666,
+      "mean_token_accuracy": 0.9572859182953835,
+      "num_tokens": 1166125.0,
+      "step": 380
+    },
+    {
+      "entropy": 0.16431492734700442,
+      "epoch": 1.388245770258237,
+      "grad_norm": 0.08056640625,
+      "learning_rate": 0.00011740689240517837,
+      "loss": 0.1597315788269043,
+      "mean_token_accuracy": 0.9589557304978371,
+      "num_tokens": 1197232.0,
+      "step": 390
+    },
+    {
+      "entropy": 0.15399955678731203,
+      "epoch": 1.4238646482635797,
+      "grad_norm": 0.07421875,
+      "learning_rate": 0.00011360838341963964,
+      "loss": 0.15647656917572023,
+      "mean_token_accuracy": 0.9591503396630288,
+      "num_tokens": 1227801.0,
+      "step": 400
+    },
+    {
+      "entropy": 0.15702628958970308,
+      "epoch": 1.4594835262689225,
+      "grad_norm": 0.0888671875,
+      "learning_rate": 0.00010978975532916189,
+      "loss": 0.16044070720672607,
+      "mean_token_accuracy": 0.958300518989563,
+      "num_tokens": 1258629.0,
+      "step": 410
+    },
+    {
+      "entropy": 0.12639004811644555,
+      "epoch": 1.4951024042742653,
+      "grad_norm": 0.0654296875,
+      "learning_rate": 0.00010595665372565027,
+      "loss": 0.12638626098632813,
+      "mean_token_accuracy": 0.9659075498580932,
+      "num_tokens": 1289072.0,
+      "step": 420
+    },
+    {
+      "entropy": 0.14171069134026765,
+      "epoch": 1.5307212822796084,
+      "grad_norm": 0.08056640625,
+      "learning_rate": 0.00010211474559915233,
+      "loss": 0.13931651115417482,
+      "mean_token_accuracy": 0.9626615524291993,
+      "num_tokens": 1319747.0,
+      "step": 430
+    },
+    {
+      "entropy": 0.16094463262706996,
+      "epoch": 1.566340160284951,
+      "grad_norm": 0.07958984375,
+      "learning_rate": 9.826971095958395e-05,
+      "loss": 0.15868637561798096,
+      "mean_token_accuracy": 0.9590194016695023,
+      "num_tokens": 1350664.0,
+      "step": 440
+    },
+    {
+      "entropy": 0.15832340456545352,
+      "epoch": 1.601959038290294,
+      "grad_norm": 0.0673828125,
+      "learning_rate": 9.442723443920623e-05,
+      "loss": 0.16317789554595946,
+      "mean_token_accuracy": 0.9585595563054085,
+      "num_tokens": 1381628.0,
+      "step": 450
+    },
+    {
+      "entropy": 0.16448966227471828,
+      "epoch": 1.6375779162956365,
+      "grad_norm": 0.0751953125,
+      "learning_rate": 9.059299688826816e-05,
+      "loss": 0.16371761560440062,
+      "mean_token_accuracy": 0.9589671149849892,
+      "num_tokens": 1412506.0,
+      "step": 460
+    },
+    {
+      "entropy": 0.15921370945870877,
+      "epoch": 1.6731967943009796,
+      "grad_norm": 0.0732421875,
+      "learning_rate": 8.677266697624138e-05,
+      "loss": 0.15553101301193237,
+      "mean_token_accuracy": 0.9597976416349411,
+      "num_tokens": 1442986.0,
+      "step": 470
+    },
+    {
+      "entropy": 0.14739455822855235,
+      "epoch": 1.7088156723063224,
+      "grad_norm": 0.0751953125,
+      "learning_rate": 8.297189281106278e-05,
+      "loss": 0.14005433320999144,
+      "mean_token_accuracy": 0.9631642028689384,
+      "num_tokens": 1473444.0,
+      "step": 480
+    },
+    {
+      "entropy": 0.1320339234545827,
+      "epoch": 1.7444345503116652,
+      "grad_norm": 0.07275390625,
+      "learning_rate": 7.919629358877657e-05,
+      "loss": 0.13607435226440429,
+      "mean_token_accuracy": 0.9643332988023758,
+      "num_tokens": 1504098.0,
+      "step": 490
+    },
+    {
+      "entropy": 0.12228272054344416,
+      "epoch": 1.780053428317008,
+      "grad_norm": 0.0888671875,
+      "learning_rate": 7.54514512859201e-05,
+      "loss": 0.12318435907363892,
+      "mean_token_accuracy": 0.9668563097715378,
+      "num_tokens": 1534785.0,
+      "step": 500
+    },
+    {
+      "entropy": 0.14428242221474646,
+      "epoch": 1.8156723063223508,
+      "grad_norm": 0.07861328125,
+      "learning_rate": 7.174290240693689e-05,
+      "loss": 0.14451712369918823,
+      "mean_token_accuracy": 0.9615424692630767,
+      "num_tokens": 1565500.0,
+      "step": 510
+    },
+    {
+      "entropy": 0.1427522897720337,
+      "epoch": 1.8512911843276938,
+      "grad_norm": 0.056884765625,
+      "learning_rate": 6.807612979881661e-05,
+      "loss": 0.1388334035873413,
+      "mean_token_accuracy": 0.9621888875961304,
+      "num_tokens": 1596446.0,
+      "step": 520
+    },
+    {
+      "entropy": 0.15157688688486814,
+      "epoch": 1.8869100623330364,
+      "grad_norm": 0.08349609375,
+      "learning_rate": 6.445655454506465e-05,
+      "loss": 0.15446548461914061,
+      "mean_token_accuracy": 0.9599128782749176,
+      "num_tokens": 1627189.0,
+      "step": 530
+    },
+    {
+      "entropy": 0.1300442773848772,
+      "epoch": 1.9225289403383794,
+      "grad_norm": 0.059326171875,
+      "learning_rate": 6.0889527950984416e-05,
+      "loss": 0.1254338026046753,
+      "mean_token_accuracy": 0.9661760002374649,
+      "num_tokens": 1657858.0,
+      "step": 540
+    },
+    {
+      "entropy": 0.1557828625664115,
+      "epoch": 1.9581478183437222,
+      "grad_norm": 0.06689453125,
+      "learning_rate": 5.738032363212258e-05,
+      "loss": 0.16083067655563354,
+      "mean_token_accuracy": 0.9595108240842819,
+      "num_tokens": 1688776.0,
+      "step": 550
+    },
+    {
+      "entropy": 0.13151839561760426,
+      "epoch": 1.993766696349065,
+      "grad_norm": 0.0810546875,
+      "learning_rate": 5.3934129717573165e-05,
+      "loss": 0.12607554197311402,
+      "mean_token_accuracy": 0.966105441749096,
+      "num_tokens": 1719573.0,
+      "step": 560
+    },
+    {
+      "epoch": 2.0,
+      "eval_entropy": 0.14480842649936676,
+      "eval_loss": 0.15477769076824188,
+      "eval_mean_token_accuracy": 0.9607759993127052,
+      "eval_num_tokens": 1725024.0,
+      "eval_runtime": 35.306,
+      "eval_samples_per_second": 15.89,
+      "eval_steps_per_second": 3.994,
+      "step": 562
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 843,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 4.057603725318144e+16,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-562/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:57851b427cb3a4649963b62f1b5d8860aa50ba7ac4abbbb2ebc56acbf65bd694
+size 5649

checkpoint-843/README.md

@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-4B-Instruct-2507
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-4B-Instruct-2507
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1

checkpoint-843/adapter_config.json

@@ -0,0 +1,46 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3-4B-Instruct-2507",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.18.1",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "down_proj",
+    "up_proj",
+    "q_proj",
+    "gate_proj",
+    "k_proj",
+    "o_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

checkpoint-843/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0fef732bbd8022ce3c2bda83792e299ed5c35b1ae1dd373c9ef1b5ff2fa61c16
+size 66127776

checkpoint-843/chat_template.jinja

@@ -0,0 +1,61 @@
+{%- 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 %}
+{%- 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" %}
+        {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- 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' }}
+{%- endif %}

checkpoint-843/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:26bf1ba2744c6b96164983ec84962f35fc51660dafcac46c99a5b1ed5e5d61ea
+size 132552715

checkpoint-843/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d302a10b4bf78a5b33d836d0dba74ecc5c351791675f6c0e73639bbf88acd3df
+size 14645

checkpoint-843/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f3eaaa49ad0633eb0127e8e12aac9b83ab64a686d17ee15c6d799822cc33157
+size 1465

checkpoint-843/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
+size 11422650

checkpoint-843/tokenizer_config.json

@@ -0,0 +1,29 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "model_max_length": 1010000,
+  "pad_token": "<|im_end|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoint-843/trainer_state.json

@@ -0,0 +1,917 @@
+{
+  "best_global_step": 562,
+  "best_metric": 0.15477769076824188,
+  "best_model_checkpoint": "/app/output/bf-router-v0.5/checkpoint-562",
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 843,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.4157426357269287,
+      "epoch": 0.003561887800534283,
+      "grad_norm": 1.203125,
+      "learning_rate": 0.0,
+      "loss": 4.336674213409424,
+      "mean_token_accuracy": 0.4680640399456024,
+      "num_tokens": 3092.0,
+      "step": 1
+    },
+    {
+      "entropy": 1.5884604156017303,
+      "epoch": 0.03561887800534283,
+      "grad_norm": 0.59375,
+      "learning_rate": 6.923076923076924e-05,
+      "loss": 3.986017862955729,
+      "mean_token_accuracy": 0.4735516524977154,
+      "num_tokens": 30650.0,
+      "step": 10
+    },
+    {
+      "entropy": 2.2315500736236573,
+      "epoch": 0.07123775601068566,
+      "grad_norm": 0.26953125,
+      "learning_rate": 0.00014615384615384615,
+      "loss": 2.16544189453125,
+      "mean_token_accuracy": 0.6272616922855377,
+      "num_tokens": 61735.0,
+      "step": 20
+    },
+    {
+      "entropy": 0.8066075079143047,
+      "epoch": 0.10685663401602849,
+      "grad_norm": 0.298828125,
+      "learning_rate": 0.00019999334629824895,
+      "loss": 0.7935717582702637,
+      "mean_token_accuracy": 0.857068446278572,
+      "num_tokens": 92484.0,
+      "step": 30
+    },
+    {
+      "entropy": 0.5340980306267739,
+      "epoch": 0.14247551202137132,
+      "grad_norm": 0.10693359375,
+      "learning_rate": 0.00019987508289735716,
+      "loss": 0.5261529922485352,
+      "mean_token_accuracy": 0.9004581302404404,
+      "num_tokens": 123010.0,
+      "step": 40
+    },
+    {
+      "entropy": 0.4001643668860197,
+      "epoch": 0.17809439002671415,
+      "grad_norm": 0.08251953125,
+      "learning_rate": 0.0001996091607179287,
+      "loss": 0.3956298351287842,
+      "mean_token_accuracy": 0.9209299519658088,
+      "num_tokens": 153710.0,
+      "step": 50
+    },
+    {
+      "entropy": 0.34828933551907537,
+      "epoch": 0.21371326803205698,
+      "grad_norm": 0.10009765625,
+      "learning_rate": 0.00019919597290851538,
+      "loss": 0.3648662090301514,
+      "mean_token_accuracy": 0.9263768374919892,
+      "num_tokens": 184577.0,
+      "step": 60
+    },
+    {
+      "entropy": 0.2983129292726517,
+      "epoch": 0.2493321460373998,
+      "grad_norm": 0.08447265625,
+      "learning_rate": 0.00019863613034027224,
+      "loss": 0.29642860889434813,
+      "mean_token_accuracy": 0.9371836110949516,
+      "num_tokens": 215040.0,
+      "step": 70
+    },
+    {
+      "entropy": 0.280751570686698,
+      "epoch": 0.28495102404274264,
+      "grad_norm": 0.099609375,
+      "learning_rate": 0.00019793046070382437,
+      "loss": 0.2813178300857544,
+      "mean_token_accuracy": 0.9388988897204399,
+      "num_tokens": 245683.0,
+      "step": 80
+    },
+    {
+      "entropy": 0.2960826952010393,
+      "epoch": 0.3205699020480855,
+      "grad_norm": 0.11962890625,
+      "learning_rate": 0.00019708000728558064,
+      "loss": 0.31115410327911375,
+      "mean_token_accuracy": 0.9335825845599175,
+      "num_tokens": 276430.0,
+      "step": 90
+    },
+    {
+      "entropy": 0.28416929207742214,
+      "epoch": 0.3561887800534283,
+      "grad_norm": 0.103515625,
+      "learning_rate": 0.00019608602742530283,
+      "loss": 0.27938365936279297,
+      "mean_token_accuracy": 0.9393619552254677,
+      "num_tokens": 307296.0,
+      "step": 100
+    },
+    {
+      "entropy": 0.23417662251740695,
+      "epoch": 0.39180765805877116,
+      "grad_norm": 0.06298828125,
+      "learning_rate": 0.00019494999065721108,
+      "loss": 0.23734774589538574,
+      "mean_token_accuracy": 0.9487677246332169,
+      "num_tokens": 337962.0,
+      "step": 110
+    },
+    {
+      "entropy": 0.24019648060202597,
+      "epoch": 0.42742653606411396,
+      "grad_norm": 0.06640625,
+      "learning_rate": 0.0001936735765373737,
+      "loss": 0.245324969291687,
+      "mean_token_accuracy": 0.9457737937569618,
+      "num_tokens": 368782.0,
+      "step": 120
+    },
+    {
+      "entropy": 0.269666058011353,
+      "epoch": 0.4630454140694568,
+      "grad_norm": 0.09765625,
+      "learning_rate": 0.00019225867216059325,
+      "loss": 0.2686375617980957,
+      "mean_token_accuracy": 0.9398508563637733,
+      "num_tokens": 399329.0,
+      "step": 130
+    },
+    {
+      "entropy": 0.18328354582190515,
+      "epoch": 0.4986642920747996,
+      "grad_norm": 0.06298828125,
+      "learning_rate": 0.00019070736937046035,
+      "loss": 0.18070420026779174,
+      "mean_token_accuracy": 0.9568557634949684,
+      "num_tokens": 429759.0,
+      "step": 140
+    },
+    {
+      "entropy": 0.21165105439722537,
+      "epoch": 0.5342831700801425,
+      "grad_norm": 0.0810546875,
+      "learning_rate": 0.0001890219616666997,
+      "loss": 0.22004930973052977,
+      "mean_token_accuracy": 0.9494227185845375,
+      "num_tokens": 460409.0,
+      "step": 150
+    },
+    {
+      "entropy": 0.23916025627404452,
+      "epoch": 0.5699020480854853,
+      "grad_norm": 0.07568359375,
+      "learning_rate": 0.00018720494081438078,
+      "loss": 0.2379377841949463,
+      "mean_token_accuracy": 0.9455192387104034,
+      "num_tokens": 491344.0,
+      "step": 160
+    },
+    {
+      "entropy": 0.2238690486177802,
+      "epoch": 0.6055209260908282,
+      "grad_norm": 0.08984375,
+      "learning_rate": 0.00018525899316000608,
+      "loss": 0.2245168685913086,
+      "mean_token_accuracy": 0.9485938593745231,
+      "num_tokens": 522029.0,
+      "step": 170
+    },
+    {
+      "entropy": 0.17032606173306702,
+      "epoch": 0.641139804096171,
+      "grad_norm": 0.080078125,
+      "learning_rate": 0.00018318699565992357,
+      "loss": 0.17549347877502441,
+      "mean_token_accuracy": 0.9574712902307511,
+      "num_tokens": 552306.0,
+      "step": 180
+    },
+    {
+      "entropy": 0.21890966054052113,
+      "epoch": 0.6767586821015138,
+      "grad_norm": 0.0673828125,
+      "learning_rate": 0.00018099201162693476,
+      "loss": 0.21968293190002441,
+      "mean_token_accuracy": 0.9482985377311707,
+      "num_tokens": 583181.0,
+      "step": 190
+    },
+    {
+      "entropy": 0.20646683853119613,
+      "epoch": 0.7123775601068566,
+      "grad_norm": 0.09765625,
+      "learning_rate": 0.00017867728620138708,
+      "loss": 0.20968315601348878,
+      "mean_token_accuracy": 0.9519409075379371,
+      "num_tokens": 613818.0,
+      "step": 200
+    },
+    {
+      "entropy": 0.23000301327556372,
+      "epoch": 0.7479964381121995,
+      "grad_norm": 0.08935546875,
+      "learning_rate": 0.00017624624155344626,
+      "loss": 0.2383474349975586,
+      "mean_token_accuracy": 0.9458372846245766,
+      "num_tokens": 644576.0,
+      "step": 210
+    },
+    {
+      "entropy": 0.21420854832977057,
+      "epoch": 0.7836153161175423,
+      "grad_norm": 0.0966796875,
+      "learning_rate": 0.0001737024718236413,
+      "loss": 0.21017465591430665,
+      "mean_token_accuracy": 0.9511569887399673,
+      "num_tokens": 675564.0,
+      "step": 220
+    },
+    {
+      "entropy": 0.19280126914381981,
+      "epoch": 0.8192341941228851,
+      "grad_norm": 0.08154296875,
+      "learning_rate": 0.00017104973780916294,
+      "loss": 0.1936139941215515,
+      "mean_token_accuracy": 0.954449312388897,
+      "num_tokens": 706309.0,
+      "step": 230
+    },
+    {
+      "entropy": 0.22524890769273043,
+      "epoch": 0.8548530721282279,
+      "grad_norm": 0.10546875,
+      "learning_rate": 0.00016829196140377085,
+      "loss": 0.22950620651245118,
+      "mean_token_accuracy": 0.9476036429405212,
+      "num_tokens": 737135.0,
+      "step": 240
+    },
+    {
+      "entropy": 0.18842433094978334,
+      "epoch": 0.8904719501335708,
+      "grad_norm": 0.087890625,
+      "learning_rate": 0.00016543321979953007,
+      "loss": 0.19049547910690307,
+      "mean_token_accuracy": 0.9543813273310662,
+      "num_tokens": 767853.0,
+      "step": 250
+    },
+    {
+      "entropy": 0.1991994746029377,
+      "epoch": 0.9260908281389136,
+      "grad_norm": 0.07763671875,
+      "learning_rate": 0.00016247773945894962,
+      "loss": 0.19527161121368408,
+      "mean_token_accuracy": 0.9535173490643502,
+      "num_tokens": 798511.0,
+      "step": 260
+    },
+    {
+      "entropy": 0.18196075949817897,
+      "epoch": 0.9617097061442564,
+      "grad_norm": 0.0771484375,
+      "learning_rate": 0.00015942988986643352,
+      "loss": 0.19091761112213135,
+      "mean_token_accuracy": 0.9535432115197182,
+      "num_tokens": 829447.0,
+      "step": 270
+    },
+    {
+      "entropy": 0.18307169582694768,
+      "epoch": 0.9973285841495992,
+      "grad_norm": 0.08837890625,
+      "learning_rate": 0.00015629417706828423,
+      "loss": 0.1818631410598755,
+      "mean_token_accuracy": 0.955574706196785,
+      "num_tokens": 860142.0,
+      "step": 280
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.18207345770817276,
+      "eval_loss": 0.18149949610233307,
+      "eval_mean_token_accuracy": 0.9555717660180221,
+      "eval_num_tokens": 862512.0,
+      "eval_runtime": 35.3256,
+      "eval_samples_per_second": 15.881,
+      "eval_steps_per_second": 3.991,
+      "step": 281
+    },
+    {
+      "entropy": 0.1545313375118451,
+      "epoch": 1.0320569902048085,
+      "grad_norm": 0.064453125,
+      "learning_rate": 0.00015307523701080768,
+      "loss": 0.14675980806350708,
+      "mean_token_accuracy": 0.9623546004295349,
+      "num_tokens": 889907.0,
+      "step": 290
+    },
+    {
+      "entropy": 0.13382284864783286,
+      "epoch": 1.0676758682101515,
+      "grad_norm": 0.060546875,
+      "learning_rate": 0.00014977782868636999,
+      "loss": 0.13706474304199218,
+      "mean_token_accuracy": 0.9633408606052398,
+      "num_tokens": 920639.0,
+      "step": 300
+    },
+    {
+      "entropy": 0.16596811451017857,
+      "epoch": 1.1032947462154943,
+      "grad_norm": 0.0791015625,
+      "learning_rate": 0.00014640682709753832,
+      "loss": 0.16318607330322266,
+      "mean_token_accuracy": 0.9578790530562401,
+      "num_tokens": 951323.0,
+      "step": 310
+    },
+    {
+      "entropy": 0.1480622159317136,
+      "epoch": 1.138913624220837,
+      "grad_norm": 0.057861328125,
+      "learning_rate": 0.0001429672160497085,
+      "loss": 0.1488279938697815,
+      "mean_token_accuracy": 0.9609164595603943,
+      "num_tokens": 981911.0,
+      "step": 320
+    },
+    {
+      "entropy": 0.1655621325597167,
+      "epoch": 1.1745325022261799,
+      "grad_norm": 0.07275390625,
+      "learning_rate": 0.00013946408078287462,
+      "loss": 0.16427644491195678,
+      "mean_token_accuracy": 0.9576473370194435,
+      "num_tokens": 1012535.0,
+      "step": 330
+    },
+    {
+      "entropy": 0.18122829273343086,
+      "epoch": 1.2101513802315227,
+      "grad_norm": 0.06689453125,
+      "learning_rate": 0.00013590260045343432,
+      "loss": 0.176006543636322,
+      "mean_token_accuracy": 0.9551453411579132,
+      "num_tokens": 1043503.0,
+      "step": 340
+    },
+    {
+      "entropy": 0.15644419118762015,
+      "epoch": 1.2457702582368655,
+      "grad_norm": 0.06201171875,
+      "learning_rate": 0.00013228804047714463,
+      "loss": 0.1652477502822876,
+      "mean_token_accuracy": 0.9575047269463539,
+      "num_tokens": 1074168.0,
+      "step": 350
+    },
+    {
+      "entropy": 0.14828295167535543,
+      "epoch": 1.2813891362422083,
+      "grad_norm": 0.06982421875,
+      "learning_rate": 0.00012862574474454928,
+      "loss": 0.14066768884658815,
+      "mean_token_accuracy": 0.9639188721776009,
+      "num_tokens": 1104681.0,
+      "step": 360
+    },
+    {
+      "entropy": 0.19175196047872306,
+      "epoch": 1.317008014247551,
+      "grad_norm": 0.09619140625,
+      "learning_rate": 0.0001249211277203859,
+      "loss": 0.1917089343070984,
+      "mean_token_accuracy": 0.9522068575024605,
+      "num_tokens": 1135361.0,
+      "step": 370
+    },
+    {
+      "entropy": 0.1644950734451413,
+      "epoch": 1.3526268922528941,
+      "grad_norm": 0.0849609375,
+      "learning_rate": 0.00012117966643865398,
+      "loss": 0.16864393949508666,
+      "mean_token_accuracy": 0.9572859182953835,
+      "num_tokens": 1166125.0,
+      "step": 380
+    },
+    {
+      "entropy": 0.16431492734700442,
+      "epoch": 1.388245770258237,
+      "grad_norm": 0.08056640625,
+      "learning_rate": 0.00011740689240517837,
+      "loss": 0.1597315788269043,
+      "mean_token_accuracy": 0.9589557304978371,
+      "num_tokens": 1197232.0,
+      "step": 390
+    },
+    {
+      "entropy": 0.15399955678731203,
+      "epoch": 1.4238646482635797,
+      "grad_norm": 0.07421875,
+      "learning_rate": 0.00011360838341963964,
+      "loss": 0.15647656917572023,
+      "mean_token_accuracy": 0.9591503396630288,
+      "num_tokens": 1227801.0,
+      "step": 400
+    },
+    {
+      "entropy": 0.15702628958970308,
+      "epoch": 1.4594835262689225,
+      "grad_norm": 0.0888671875,
+      "learning_rate": 0.00010978975532916189,
+      "loss": 0.16044070720672607,
+      "mean_token_accuracy": 0.958300518989563,
+      "num_tokens": 1258629.0,
+      "step": 410
+    },
+    {
+      "entropy": 0.12639004811644555,
+      "epoch": 1.4951024042742653,
+      "grad_norm": 0.0654296875,
+      "learning_rate": 0.00010595665372565027,
+      "loss": 0.12638626098632813,
+      "mean_token_accuracy": 0.9659075498580932,
+      "num_tokens": 1289072.0,
+      "step": 420
+    },
+    {
+      "entropy": 0.14171069134026765,
+      "epoch": 1.5307212822796084,
+      "grad_norm": 0.08056640625,
+      "learning_rate": 0.00010211474559915233,
+      "loss": 0.13931651115417482,
+      "mean_token_accuracy": 0.9626615524291993,
+      "num_tokens": 1319747.0,
+      "step": 430
+    },
+    {
+      "entropy": 0.16094463262706996,
+      "epoch": 1.566340160284951,
+      "grad_norm": 0.07958984375,
+      "learning_rate": 9.826971095958395e-05,
+      "loss": 0.15868637561798096,
+      "mean_token_accuracy": 0.9590194016695023,
+      "num_tokens": 1350664.0,
+      "step": 440
+    },
+    {
+      "entropy": 0.15832340456545352,
+      "epoch": 1.601959038290294,
+      "grad_norm": 0.0673828125,
+      "learning_rate": 9.442723443920623e-05,
+      "loss": 0.16317789554595946,
+      "mean_token_accuracy": 0.9585595563054085,
+      "num_tokens": 1381628.0,
+      "step": 450
+    },
+    {
+      "entropy": 0.16448966227471828,
+      "epoch": 1.6375779162956365,
+      "grad_norm": 0.0751953125,
+      "learning_rate": 9.059299688826816e-05,
+      "loss": 0.16371761560440062,
+      "mean_token_accuracy": 0.9589671149849892,
+      "num_tokens": 1412506.0,
+      "step": 460
+    },
+    {
+      "entropy": 0.15921370945870877,
+      "epoch": 1.6731967943009796,
+      "grad_norm": 0.0732421875,
+      "learning_rate": 8.677266697624138e-05,
+      "loss": 0.15553101301193237,
+      "mean_token_accuracy": 0.9597976416349411,
+      "num_tokens": 1442986.0,
+      "step": 470
+    },
+    {
+      "entropy": 0.14739455822855235,
+      "epoch": 1.7088156723063224,
+      "grad_norm": 0.0751953125,
+      "learning_rate": 8.297189281106278e-05,
+      "loss": 0.14005433320999144,
+      "mean_token_accuracy": 0.9631642028689384,
+      "num_tokens": 1473444.0,
+      "step": 480
+    },
+    {
+      "entropy": 0.1320339234545827,
+      "epoch": 1.7444345503116652,
+      "grad_norm": 0.07275390625,
+      "learning_rate": 7.919629358877657e-05,
+      "loss": 0.13607435226440429,
+      "mean_token_accuracy": 0.9643332988023758,
+      "num_tokens": 1504098.0,
+      "step": 490
+    },
+    {
+      "entropy": 0.12228272054344416,
+      "epoch": 1.780053428317008,
+      "grad_norm": 0.0888671875,
+      "learning_rate": 7.54514512859201e-05,
+      "loss": 0.12318435907363892,
+      "mean_token_accuracy": 0.9668563097715378,
+      "num_tokens": 1534785.0,
+      "step": 500
+    },
+    {
+      "entropy": 0.14428242221474646,
+      "epoch": 1.8156723063223508,
+      "grad_norm": 0.07861328125,
+      "learning_rate": 7.174290240693689e-05,
+      "loss": 0.14451712369918823,
+      "mean_token_accuracy": 0.9615424692630767,
+      "num_tokens": 1565500.0,
+      "step": 510
+    },
+    {
+      "entropy": 0.1427522897720337,
+      "epoch": 1.8512911843276938,
+      "grad_norm": 0.056884765625,
+      "learning_rate": 6.807612979881661e-05,
+      "loss": 0.1388334035873413,
+      "mean_token_accuracy": 0.9621888875961304,
+      "num_tokens": 1596446.0,
+      "step": 520
+    },
+    {
+      "entropy": 0.15157688688486814,
+      "epoch": 1.8869100623330364,
+      "grad_norm": 0.08349609375,
+      "learning_rate": 6.445655454506465e-05,
+      "loss": 0.15446548461914061,
+      "mean_token_accuracy": 0.9599128782749176,
+      "num_tokens": 1627189.0,
+      "step": 530
+    },
+    {
+      "entropy": 0.1300442773848772,
+      "epoch": 1.9225289403383794,
+      "grad_norm": 0.059326171875,
+      "learning_rate": 6.0889527950984416e-05,
+      "loss": 0.1254338026046753,
+      "mean_token_accuracy": 0.9661760002374649,
+      "num_tokens": 1657858.0,
+      "step": 540
+    },
+    {
+      "entropy": 0.1557828625664115,
+      "epoch": 1.9581478183437222,
+      "grad_norm": 0.06689453125,
+      "learning_rate": 5.738032363212258e-05,
+      "loss": 0.16083067655563354,
+      "mean_token_accuracy": 0.9595108240842819,
+      "num_tokens": 1688776.0,
+      "step": 550
+    },
+    {
+      "entropy": 0.13151839561760426,
+      "epoch": 1.993766696349065,
+      "grad_norm": 0.0810546875,
+      "learning_rate": 5.3934129717573165e-05,
+      "loss": 0.12607554197311402,
+      "mean_token_accuracy": 0.966105441749096,
+      "num_tokens": 1719573.0,
+      "step": 560
+    },
+    {
+      "epoch": 2.0,
+      "eval_entropy": 0.14480842649936676,
+      "eval_loss": 0.15477769076824188,
+      "eval_mean_token_accuracy": 0.9607759993127052,
+      "eval_num_tokens": 1725024.0,
+      "eval_runtime": 35.306,
+      "eval_samples_per_second": 15.89,
+      "eval_steps_per_second": 3.994,
+      "step": 562
+    },
+    {
+      "entropy": 0.13348992398151985,
+      "epoch": 2.0284951024042743,
+      "grad_norm": 0.09326171875,
+      "learning_rate": 5.0556041179668354e-05,
+      "loss": 0.12256227731704712,
+      "mean_token_accuracy": 0.9662964267608447,
+      "num_tokens": 1749504.0,
+      "step": 570
+    },
+    {
+      "entropy": 0.13551258575171232,
+      "epoch": 2.064113980409617,
+      "grad_norm": 0.0732421875,
+      "learning_rate": 4.725105230139465e-05,
+      "loss": 0.12811031341552734,
+      "mean_token_accuracy": 0.9640967190265656,
+      "num_tokens": 1780206.0,
+      "step": 580
+    },
+    {
+      "entropy": 0.11903938055038452,
+      "epoch": 2.09973285841496,
+      "grad_norm": 0.061279296875,
+      "learning_rate": 4.402404929267235e-05,
+      "loss": 0.1129370927810669,
+      "mean_token_accuracy": 0.9682316735386849,
+      "num_tokens": 1810985.0,
+      "step": 590
+    },
+    {
+      "entropy": 0.11653038449585437,
+      "epoch": 2.135351736420303,
+      "grad_norm": 0.08056640625,
+      "learning_rate": 4.087980306641365e-05,
+      "loss": 0.11388391256332397,
+      "mean_token_accuracy": 0.9668685078620911,
+      "num_tokens": 1841597.0,
+      "step": 600
+    },
+    {
+      "entropy": 0.11204307321459055,
+      "epoch": 2.1709706144256455,
+      "grad_norm": 0.1015625,
+      "learning_rate": 3.7822962185039914e-05,
+      "loss": 0.11144263744354248,
+      "mean_token_accuracy": 0.9687378108501434,
+      "num_tokens": 1872199.0,
+      "step": 610
+    },
+    {
+      "entropy": 0.12015030086040497,
+      "epoch": 2.2065894924309886,
+      "grad_norm": 0.0712890625,
+      "learning_rate": 3.4858045987886145e-05,
+      "loss": 0.11484370231628419,
+      "mean_token_accuracy": 0.967641019821167,
+      "num_tokens": 1903172.0,
+      "step": 620
+    },
+    {
+      "entropy": 0.12499598301947117,
+      "epoch": 2.242208370436331,
+      "grad_norm": 0.06640625,
+      "learning_rate": 3.198943790965332e-05,
+      "loss": 0.12357065677642823,
+      "mean_token_accuracy": 0.9670094296336174,
+      "num_tokens": 1933713.0,
+      "step": 630
+    },
+    {
+      "entropy": 0.1296376422047615,
+      "epoch": 2.277827248441674,
+      "grad_norm": 0.0673828125,
+      "learning_rate": 2.9221378999786853e-05,
+      "loss": 0.12340972423553467,
+      "mean_token_accuracy": 0.96511862128973,
+      "num_tokens": 1964556.0,
+      "step": 640
+    },
+    {
+      "entropy": 0.12314727567136288,
+      "epoch": 2.3134461264470167,
+      "grad_norm": 0.080078125,
+      "learning_rate": 2.655796165236234e-05,
+      "loss": 0.11837244033813477,
+      "mean_token_accuracy": 0.9671967878937722,
+      "num_tokens": 1995252.0,
+      "step": 650
+    },
+    {
+      "entropy": 0.11561896577477455,
+      "epoch": 2.3490650044523598,
+      "grad_norm": 0.1064453125,
+      "learning_rate": 2.400312355574853e-05,
+      "loss": 0.11029613018035889,
+      "mean_token_accuracy": 0.967219403386116,
+      "num_tokens": 2025889.0,
+      "step": 660
+    },
+    {
+      "entropy": 0.11878675390034914,
+      "epoch": 2.3846838824577024,
+      "grad_norm": 0.080078125,
+      "learning_rate": 2.1560641870992616e-05,
+      "loss": 0.11299164295196533,
+      "mean_token_accuracy": 0.9675500631332398,
+      "num_tokens": 2056588.0,
+      "step": 670
+    },
+    {
+      "entropy": 0.11508079580962657,
+      "epoch": 2.4203027604630454,
+      "grad_norm": 0.07080078125,
+      "learning_rate": 1.9234127647534604e-05,
+      "loss": 0.10480635166168213,
+      "mean_token_accuracy": 0.9691305905580521,
+      "num_tokens": 2087259.0,
+      "step": 680
+    },
+    {
+      "entropy": 0.11411711536347866,
+      "epoch": 2.4559216384683884,
+      "grad_norm": 0.06640625,
+      "learning_rate": 1.7027020484506996e-05,
+      "loss": 0.11348260641098022,
+      "mean_token_accuracy": 0.9682735517621041,
+      "num_tokens": 2117728.0,
+      "step": 690
+    },
+    {
+      "entropy": 0.11259160749614239,
+      "epoch": 2.491540516473731,
+      "grad_norm": 0.05224609375,
+      "learning_rate": 1.4942583445512103e-05,
+      "loss": 0.10898959636688232,
+      "mean_token_accuracy": 0.9685835257172585,
+      "num_tokens": 2148333.0,
+      "step": 700
+    },
+    {
+      "entropy": 0.12499043270945549,
+      "epoch": 2.527159394479074,
+      "grad_norm": 0.0849609375,
+      "learning_rate": 1.2983898234396308e-05,
+      "loss": 0.12133818864822388,
+      "mean_token_accuracy": 0.9664898782968521,
+      "num_tokens": 2178932.0,
+      "step": 710
+    },
+    {
+      "entropy": 0.11083998121321201,
+      "epoch": 2.5627782724844166,
+      "grad_norm": 0.048828125,
+      "learning_rate": 1.1153860639152169e-05,
+      "loss": 0.10544953346252442,
+      "mean_token_accuracy": 0.9696028590202331,
+      "num_tokens": 2209610.0,
+      "step": 720
+    },
+    {
+      "entropy": 0.11835917010903359,
+      "epoch": 2.5983971504897596,
+      "grad_norm": 0.08203125,
+      "learning_rate": 9.455176250685338e-06,
+      "loss": 0.11609755754470825,
+      "mean_token_accuracy": 0.9664707094430923,
+      "num_tokens": 2240346.0,
+      "step": 730
+    },
+    {
+      "entropy": 0.12215895913541316,
+      "epoch": 2.634016028495102,
+      "grad_norm": 0.0673828125,
+      "learning_rate": 7.890356462775373e-06,
+      "loss": 0.11725597381591797,
+      "mean_token_accuracy": 0.9670157313346863,
+      "num_tokens": 2271054.0,
+      "step": 740
+    },
+    {
+      "entropy": 0.12366384714841842,
+      "epoch": 2.669634906500445,
+      "grad_norm": 0.07861328125,
+      "learning_rate": 6.461714759144233e-06,
+      "loss": 0.1168657660484314,
+      "mean_token_accuracy": 0.9668171271681786,
+      "num_tokens": 2301971.0,
+      "step": 750
+    },
+    {
+      "entropy": 0.10823128782212735,
+      "epoch": 2.7052537845057882,
+      "grad_norm": 0.05712890625,
+      "learning_rate": 5.171363293121901e-06,
+      "loss": 0.10103480815887451,
+      "mean_token_accuracy": 0.9696328729391098,
+      "num_tokens": 2332515.0,
+      "step": 760
+    },
+    {
+      "entropy": 0.10126303266733885,
+      "epoch": 2.740872662511131,
+      "grad_norm": 0.07080078125,
+      "learning_rate": 4.021209764965828e-06,
+      "loss": 0.09550071954727173,
+      "mean_token_accuracy": 0.9718543246388436,
+      "num_tokens": 2363176.0,
+      "step": 770
+    },
+    {
+      "entropy": 0.10594012532383204,
+      "epoch": 2.776491540516474,
+      "grad_norm": 0.0625,
+      "learning_rate": 3.0129546014508567e-06,
+      "loss": 0.1027148962020874,
+      "mean_token_accuracy": 0.9698196157813073,
+      "num_tokens": 2393823.0,
+      "step": 780
+    },
+    {
+      "entropy": 0.11277557052671909,
+      "epoch": 2.8121104185218164,
+      "grad_norm": 0.057373046875,
+      "learning_rate": 2.1480884418993983e-06,
+      "loss": 0.10367020368576049,
+      "mean_token_accuracy": 0.9688401147723198,
+      "num_tokens": 2424575.0,
+      "step": 790
+    },
+    {
+      "entropy": 0.13837510608136655,
+      "epoch": 2.8477292965271594,
+      "grad_norm": 0.0791015625,
+      "learning_rate": 1.4278899343687425e-06,
+      "loss": 0.13071552515029908,
+      "mean_token_accuracy": 0.9625913232564927,
+      "num_tokens": 2455901.0,
+      "step": 800
+    },
+    {
+      "entropy": 0.12289627455174923,
+      "epoch": 2.883348174532502,
+      "grad_norm": 0.09423828125,
+      "learning_rate": 8.534238452534759e-07,
+      "loss": 0.11349356174468994,
+      "mean_token_accuracy": 0.9673542976379395,
+      "num_tokens": 2486966.0,
+      "step": 810
+    },
+    {
+      "entropy": 0.11963035948574544,
+      "epoch": 2.918967052537845,
+      "grad_norm": 0.05419921875,
+      "learning_rate": 4.2553948509802545e-07,
+      "loss": 0.11257882118225097,
+      "mean_token_accuracy": 0.9673726871609688,
+      "num_tokens": 2517590.0,
+      "step": 820
+    },
+    {
+      "entropy": 0.13870051857084037,
+      "epoch": 2.954585930543188,
+      "grad_norm": 0.06298828125,
+      "learning_rate": 1.4486945294652776e-07,
+      "loss": 0.13519666194915772,
+      "mean_token_accuracy": 0.9639795452356339,
+      "num_tokens": 2548330.0,
+      "step": 830
+    },
+    {
+      "entropy": 0.10831217430531978,
+      "epoch": 2.9902048085485307,
+      "grad_norm": 0.06494140625,
+      "learning_rate": 1.1828701086558092e-08,
+      "loss": 0.10172584056854247,
+      "mean_token_accuracy": 0.9707534283399581,
+      "num_tokens": 2579075.0,
+      "step": 840
+    },
+    {
+      "epoch": 3.0,
+      "eval_entropy": 0.12990308695651115,
+      "eval_loss": 0.15479549765586853,
+      "eval_mean_token_accuracy": 0.9612296595641062,
+      "eval_num_tokens": 2587536.0,
+      "eval_runtime": 35.3111,
+      "eval_samples_per_second": 15.887,
+      "eval_steps_per_second": 3.993,
+      "step": 843
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 843,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "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": 6.085644413841408e+16,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-843/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:57851b427cb3a4649963b62f1b5d8860aa50ba7ac4abbbb2ebc56acbf65bd694
+size 5649

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
+size 11422650

tokenizer_config.json

@@ -0,0 +1,29 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "model_max_length": 1010000,
+  "pad_token": "<|im_end|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:57851b427cb3a4649963b62f1b5d8860aa50ba7ac4abbbb2ebc56acbf65bd694
+size 5649