Upload folder using huggingface_hub

README.md

@@ -0,0 +1,202 @@
+---
+base_model: HuggingFaceTB/SmolVLM-Instruct
+library_name: peft
+---
+
+# 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 0.11.1

adapter_config.json

@@ -0,0 +1,26 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "HuggingFaceTB/SmolVLM-Instruct",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": "gaussian",
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_dropout": 0.1,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": "(.*(model.text_model).*(down_proj|gate_proj|up_proj|k_proj|q_proj|v_proj|o_proj).*$|.*(custom_text_proj).*$)",
+  "task_type": "FEATURE_EXTRACTION",
+  "use_dora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:075ad26e6ff42609469cbc3e30a907843286eef90f1527618f9aac24b9babdd7
+size 72400408

added_tokens.json

@@ -0,0 +1,5 @@
+{
+  "<end_of_utterance>": 49154,
+  "<fake_token_around_image>": 49152,
+  "<image>": 49153
+}

chat_template.json

@@ -0,0 +1,3 @@
+{
+  "chat_template": "<|im_start|>{% for message in messages %}{{message['role'].capitalize()}}{% if message['content'][0]['type'] == 'image' %}{{':'}}{% else %}{{': '}}{% endif %}{% for line in message['content'] %}{% if line['type'] == 'text' %}{{line['text']}}{% elif line['type'] == 'image' %}{{ '<image>' }}{% endif %}{% endfor %}<end_of_utterance>\n{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}"
+}

checkpoint-2772/README.md

@@ -0,0 +1,202 @@
+---
+base_model: ./models/SmolVLM-Instruct
+library_name: peft
+---
+
+# 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 0.11.1

checkpoint-2772/adapter_config.json

@@ -0,0 +1,26 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "./models/SmolVLM-Instruct",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": "gaussian",
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_dropout": 0.1,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": "(.*(model.text_model).*(down_proj|gate_proj|up_proj|k_proj|q_proj|v_proj|o_proj).*$|.*(custom_text_proj).*$)",
+  "task_type": "FEATURE_EXTRACTION",
+  "use_dora": false,
+  "use_rslora": false
+}

checkpoint-2772/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:075ad26e6ff42609469cbc3e30a907843286eef90f1527618f9aac24b9babdd7
+size 72400408

checkpoint-2772/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d2129dcfb6562f7bc6bdad9fe51396193346aec5da9eba61402f90832898dede
+size 144986754

checkpoint-2772/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3d8db28d075b6611d668679b2d500db5ef1992227613ebb7246a23d8216db78
+size 14244

checkpoint-2772/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc512ccb23c53760787387696789fcd4ddf0883dd39c3a82a3e4c7906bbec9b0
+size 1064

checkpoint-2772/trainer_state.json

@@ -0,0 +1,2188 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 100,
+  "global_step": 2772,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.010822510822510822,
+      "grad_norm": 1.453125,
+      "learning_rate": 5e-05,
+      "loss": 1.0068,
+      "step": 10
+    },
+    {
+      "epoch": 0.021645021645021644,
+      "grad_norm": 0.7421875,
+      "learning_rate": 0.0001,
+      "loss": 0.7642,
+      "step": 20
+    },
+    {
+      "epoch": 0.032467532467532464,
+      "grad_norm": 1.140625,
+      "learning_rate": 0.00015,
+      "loss": 0.676,
+      "step": 30
+    },
+    {
+      "epoch": 0.04329004329004329,
+      "grad_norm": 0.2890625,
+      "learning_rate": 0.0002,
+      "loss": 0.6209,
+      "step": 40
+    },
+    {
+      "epoch": 0.05411255411255411,
+      "grad_norm": 0.4453125,
+      "learning_rate": 0.00025,
+      "loss": 0.5199,
+      "step": 50
+    },
+    {
+      "epoch": 0.06493506493506493,
+      "grad_norm": 0.353515625,
+      "learning_rate": 0.0003,
+      "loss": 0.3979,
+      "step": 60
+    },
+    {
+      "epoch": 0.07575757575757576,
+      "grad_norm": 0.80859375,
+      "learning_rate": 0.00035,
+      "loss": 0.3112,
+      "step": 70
+    },
+    {
+      "epoch": 0.08658008658008658,
+      "grad_norm": 0.56640625,
+      "learning_rate": 0.0004,
+      "loss": 0.2598,
+      "step": 80
+    },
+    {
+      "epoch": 0.09740259740259741,
+      "grad_norm": 0.25,
+      "learning_rate": 0.00045000000000000004,
+      "loss": 0.2816,
+      "step": 90
+    },
+    {
+      "epoch": 0.10822510822510822,
+      "grad_norm": 0.359375,
+      "learning_rate": 0.0005,
+      "loss": 0.2459,
+      "step": 100
+    },
+    {
+      "epoch": 0.10822510822510822,
+      "eval_loss": 0.2404150664806366,
+      "eval_runtime": 62.0434,
+      "eval_samples_per_second": 8.059,
+      "eval_steps_per_second": 0.064,
+      "step": 100
+    },
+    {
+      "epoch": 0.11904761904761904,
+      "grad_norm": 0.447265625,
+      "learning_rate": 0.0004981287425149701,
+      "loss": 0.2341,
+      "step": 110
+    },
+    {
+      "epoch": 0.12987012987012986,
+      "grad_norm": 0.353515625,
+      "learning_rate": 0.0004962574850299401,
+      "loss": 0.2285,
+      "step": 120
+    },
+    {
+      "epoch": 0.1406926406926407,
+      "grad_norm": 0.314453125,
+      "learning_rate": 0.0004943862275449102,
+      "loss": 0.2229,
+      "step": 130
+    },
+    {
+      "epoch": 0.15151515151515152,
+      "grad_norm": 0.392578125,
+      "learning_rate": 0.0004925149700598803,
+      "loss": 0.2184,
+      "step": 140
+    },
+    {
+      "epoch": 0.16233766233766234,
+      "grad_norm": 0.263671875,
+      "learning_rate": 0.0004906437125748503,
+      "loss": 0.2225,
+      "step": 150
+    },
+    {
+      "epoch": 0.17316017316017315,
+      "grad_norm": 0.474609375,
+      "learning_rate": 0.0004887724550898204,
+      "loss": 0.2258,
+      "step": 160
+    },
+    {
+      "epoch": 0.18398268398268397,
+      "grad_norm": 0.408203125,
+      "learning_rate": 0.0004869011976047904,
+      "loss": 0.2228,
+      "step": 170
+    },
+    {
+      "epoch": 0.19480519480519481,
+      "grad_norm": 0.2734375,
+      "learning_rate": 0.00048502994011976046,
+      "loss": 0.2226,
+      "step": 180
+    },
+    {
+      "epoch": 0.20562770562770563,
+      "grad_norm": 0.236328125,
+      "learning_rate": 0.00048315868263473056,
+      "loss": 0.2105,
+      "step": 190
+    },
+    {
+      "epoch": 0.21645021645021645,
+      "grad_norm": 0.3125,
+      "learning_rate": 0.0004812874251497006,
+      "loss": 0.1944,
+      "step": 200
+    },
+    {
+      "epoch": 0.21645021645021645,
+      "eval_loss": 0.21168169379234314,
+      "eval_runtime": 46.9599,
+      "eval_samples_per_second": 10.647,
+      "eval_steps_per_second": 0.085,
+      "step": 200
+    },
+    {
+      "epoch": 0.22727272727272727,
+      "grad_norm": 0.29296875,
+      "learning_rate": 0.00047941616766467065,
+      "loss": 0.177,
+      "step": 210
+    },
+    {
+      "epoch": 0.23809523809523808,
+      "grad_norm": 0.203125,
+      "learning_rate": 0.00047754491017964074,
+      "loss": 0.2048,
+      "step": 220
+    },
+    {
+      "epoch": 0.24891774891774893,
+      "grad_norm": 0.32421875,
+      "learning_rate": 0.0004756736526946108,
+      "loss": 0.185,
+      "step": 230
+    },
+    {
+      "epoch": 0.2597402597402597,
+      "grad_norm": 0.279296875,
+      "learning_rate": 0.00047380239520958083,
+      "loss": 0.1941,
+      "step": 240
+    },
+    {
+      "epoch": 0.27056277056277056,
+      "grad_norm": 0.30078125,
+      "learning_rate": 0.0004719311377245509,
+      "loss": 0.1975,
+      "step": 250
+    },
+    {
+      "epoch": 0.2813852813852814,
+      "grad_norm": 0.2197265625,
+      "learning_rate": 0.00047005988023952097,
+      "loss": 0.1745,
+      "step": 260
+    },
+    {
+      "epoch": 0.2922077922077922,
+      "grad_norm": 0.28515625,
+      "learning_rate": 0.000468188622754491,
+      "loss": 0.1835,
+      "step": 270
+    },
+    {
+      "epoch": 0.30303030303030304,
+      "grad_norm": 0.2177734375,
+      "learning_rate": 0.0004663173652694611,
+      "loss": 0.2128,
+      "step": 280
+    },
+    {
+      "epoch": 0.31385281385281383,
+      "grad_norm": 0.2490234375,
+      "learning_rate": 0.00046444610778443115,
+      "loss": 0.1647,
+      "step": 290
+    },
+    {
+      "epoch": 0.3246753246753247,
+      "grad_norm": 0.2412109375,
+      "learning_rate": 0.0004625748502994012,
+      "loss": 0.1947,
+      "step": 300
+    },
+    {
+      "epoch": 0.3246753246753247,
+      "eval_loss": 0.18798980116844177,
+      "eval_runtime": 47.046,
+      "eval_samples_per_second": 10.628,
+      "eval_steps_per_second": 0.085,
+      "step": 300
+    },
+    {
+      "epoch": 0.3354978354978355,
+      "grad_norm": 0.2392578125,
+      "learning_rate": 0.00046070359281437124,
+      "loss": 0.2069,
+      "step": 310
+    },
+    {
+      "epoch": 0.3463203463203463,
+      "grad_norm": 0.26953125,
+      "learning_rate": 0.00045883233532934134,
+      "loss": 0.2029,
+      "step": 320
+    },
+    {
+      "epoch": 0.35714285714285715,
+      "grad_norm": 0.26171875,
+      "learning_rate": 0.0004569610778443114,
+      "loss": 0.2126,
+      "step": 330
+    },
+    {
+      "epoch": 0.36796536796536794,
+      "grad_norm": 0.16015625,
+      "learning_rate": 0.0004550898203592814,
+      "loss": 0.1838,
+      "step": 340
+    },
+    {
+      "epoch": 0.3787878787878788,
+      "grad_norm": 0.271484375,
+      "learning_rate": 0.0004532185628742515,
+      "loss": 0.1869,
+      "step": 350
+    },
+    {
+      "epoch": 0.38961038961038963,
+      "grad_norm": 0.2333984375,
+      "learning_rate": 0.00045134730538922156,
+      "loss": 0.1897,
+      "step": 360
+    },
+    {
+      "epoch": 0.4004329004329004,
+      "grad_norm": 0.2236328125,
+      "learning_rate": 0.0004494760479041916,
+      "loss": 0.1896,
+      "step": 370
+    },
+    {
+      "epoch": 0.41125541125541126,
+      "grad_norm": 0.275390625,
+      "learning_rate": 0.0004476047904191617,
+      "loss": 0.1956,
+      "step": 380
+    },
+    {
+      "epoch": 0.42207792207792205,
+      "grad_norm": 0.2099609375,
+      "learning_rate": 0.00044573353293413174,
+      "loss": 0.1776,
+      "step": 390
+    },
+    {
+      "epoch": 0.4329004329004329,
+      "grad_norm": 0.2314453125,
+      "learning_rate": 0.0004438622754491018,
+      "loss": 0.1866,
+      "step": 400
+    },
+    {
+      "epoch": 0.4329004329004329,
+      "eval_loss": 0.1857587993144989,
+      "eval_runtime": 49.3784,
+      "eval_samples_per_second": 10.126,
+      "eval_steps_per_second": 0.081,
+      "step": 400
+    },
+    {
+      "epoch": 0.44372294372294374,
+      "grad_norm": 0.37109375,
+      "learning_rate": 0.0004419910179640719,
+      "loss": 0.1785,
+      "step": 410
+    },
+    {
+      "epoch": 0.45454545454545453,
+      "grad_norm": 0.26953125,
+      "learning_rate": 0.00044011976047904193,
+      "loss": 0.2015,
+      "step": 420
+    },
+    {
+      "epoch": 0.4653679653679654,
+      "grad_norm": 0.26171875,
+      "learning_rate": 0.00043824850299401197,
+      "loss": 0.1682,
+      "step": 430
+    },
+    {
+      "epoch": 0.47619047619047616,
+      "grad_norm": 0.2734375,
+      "learning_rate": 0.00043637724550898207,
+      "loss": 0.1783,
+      "step": 440
+    },
+    {
+      "epoch": 0.487012987012987,
+      "grad_norm": 0.2138671875,
+      "learning_rate": 0.00043450598802395206,
+      "loss": 0.188,
+      "step": 450
+    },
+    {
+      "epoch": 0.49783549783549785,
+      "grad_norm": 0.1669921875,
+      "learning_rate": 0.00043263473053892215,
+      "loss": 0.1804,
+      "step": 460
+    },
+    {
+      "epoch": 0.5086580086580087,
+      "grad_norm": 0.26953125,
+      "learning_rate": 0.00043076347305389225,
+      "loss": 0.1698,
+      "step": 470
+    },
+    {
+      "epoch": 0.5194805194805194,
+      "grad_norm": 0.291015625,
+      "learning_rate": 0.00042889221556886224,
+      "loss": 0.1776,
+      "step": 480
+    },
+    {
+      "epoch": 0.5303030303030303,
+      "grad_norm": 0.287109375,
+      "learning_rate": 0.00042702095808383234,
+      "loss": 0.1753,
+      "step": 490
+    },
+    {
+      "epoch": 0.5411255411255411,
+      "grad_norm": 0.244140625,
+      "learning_rate": 0.00042514970059880243,
+      "loss": 0.1786,
+      "step": 500
+    },
+    {
+      "epoch": 0.5411255411255411,
+      "eval_loss": 0.18613742291927338,
+      "eval_runtime": 49.4353,
+      "eval_samples_per_second": 10.114,
+      "eval_steps_per_second": 0.081,
+      "step": 500
+    },
+    {
+      "epoch": 0.551948051948052,
+      "grad_norm": 0.30859375,
+      "learning_rate": 0.0004232784431137724,
+      "loss": 0.1558,
+      "step": 510
+    },
+    {
+      "epoch": 0.5627705627705628,
+      "grad_norm": 0.1484375,
+      "learning_rate": 0.0004214071856287425,
+      "loss": 0.1595,
+      "step": 520
+    },
+    {
+      "epoch": 0.5735930735930735,
+      "grad_norm": 0.208984375,
+      "learning_rate": 0.0004195359281437126,
+      "loss": 0.161,
+      "step": 530
+    },
+    {
+      "epoch": 0.5844155844155844,
+      "grad_norm": 0.21484375,
+      "learning_rate": 0.00041766467065868266,
+      "loss": 0.1699,
+      "step": 540
+    },
+    {
+      "epoch": 0.5952380952380952,
+      "grad_norm": 0.2890625,
+      "learning_rate": 0.0004157934131736527,
+      "loss": 0.1694,
+      "step": 550
+    },
+    {
+      "epoch": 0.6060606060606061,
+      "grad_norm": 0.1943359375,
+      "learning_rate": 0.00041392215568862275,
+      "loss": 0.1753,
+      "step": 560
+    },
+    {
+      "epoch": 0.6168831168831169,
+      "grad_norm": 0.185546875,
+      "learning_rate": 0.00041205089820359284,
+      "loss": 0.1781,
+      "step": 570
+    },
+    {
+      "epoch": 0.6277056277056277,
+      "grad_norm": 0.283203125,
+      "learning_rate": 0.0004101796407185629,
+      "loss": 0.1778,
+      "step": 580
+    },
+    {
+      "epoch": 0.6385281385281385,
+      "grad_norm": 0.208984375,
+      "learning_rate": 0.00040830838323353293,
+      "loss": 0.1501,
+      "step": 590
+    },
+    {
+      "epoch": 0.6493506493506493,
+      "grad_norm": 0.2021484375,
+      "learning_rate": 0.000406437125748503,
+      "loss": 0.1855,
+      "step": 600
+    },
+    {
+      "epoch": 0.6493506493506493,
+      "eval_loss": 0.18048742413520813,
+      "eval_runtime": 52.6384,
+      "eval_samples_per_second": 9.499,
+      "eval_steps_per_second": 0.076,
+      "step": 600
+    },
+    {
+      "epoch": 0.6601731601731602,
+      "grad_norm": 0.162109375,
+      "learning_rate": 0.00040456586826347307,
+      "loss": 0.1783,
+      "step": 610
+    },
+    {
+      "epoch": 0.670995670995671,
+      "grad_norm": 0.330078125,
+      "learning_rate": 0.0004026946107784431,
+      "loss": 0.1604,
+      "step": 620
+    },
+    {
+      "epoch": 0.6818181818181818,
+      "grad_norm": 0.205078125,
+      "learning_rate": 0.0004008233532934132,
+      "loss": 0.168,
+      "step": 630
+    },
+    {
+      "epoch": 0.6926406926406926,
+      "grad_norm": 0.2080078125,
+      "learning_rate": 0.00039895209580838325,
+      "loss": 0.1583,
+      "step": 640
+    },
+    {
+      "epoch": 0.7034632034632035,
+      "grad_norm": 0.203125,
+      "learning_rate": 0.0003970808383233533,
+      "loss": 0.1432,
+      "step": 650
+    },
+    {
+      "epoch": 0.7142857142857143,
+      "grad_norm": 0.2177734375,
+      "learning_rate": 0.0003952095808383234,
+      "loss": 0.1823,
+      "step": 660
+    },
+    {
+      "epoch": 0.7251082251082251,
+      "grad_norm": 0.2275390625,
+      "learning_rate": 0.00039333832335329343,
+      "loss": 0.1776,
+      "step": 670
+    },
+    {
+      "epoch": 0.7359307359307359,
+      "grad_norm": 0.234375,
+      "learning_rate": 0.0003914670658682635,
+      "loss": 0.1513,
+      "step": 680
+    },
+    {
+      "epoch": 0.7467532467532467,
+      "grad_norm": 0.283203125,
+      "learning_rate": 0.0003895958083832336,
+      "loss": 0.1692,
+      "step": 690
+    },
+    {
+      "epoch": 0.7575757575757576,
+      "grad_norm": 0.140625,
+      "learning_rate": 0.00038772455089820356,
+      "loss": 0.1716,
+      "step": 700
+    },
+    {
+      "epoch": 0.7575757575757576,
+      "eval_loss": 0.16686995327472687,
+      "eval_runtime": 48.5544,
+      "eval_samples_per_second": 10.298,
+      "eval_steps_per_second": 0.082,
+      "step": 700
+    },
+    {
+      "epoch": 0.7683982683982684,
+      "grad_norm": 0.28125,
+      "learning_rate": 0.00038585329341317366,
+      "loss": 0.1619,
+      "step": 710
+    },
+    {
+      "epoch": 0.7792207792207793,
+      "grad_norm": 0.2197265625,
+      "learning_rate": 0.00038398203592814376,
+      "loss": 0.167,
+      "step": 720
+    },
+    {
+      "epoch": 0.79004329004329,
+      "grad_norm": 0.3125,
+      "learning_rate": 0.00038211077844311375,
+      "loss": 0.1507,
+      "step": 730
+    },
+    {
+      "epoch": 0.8008658008658008,
+      "grad_norm": 0.390625,
+      "learning_rate": 0.00038023952095808384,
+      "loss": 0.1649,
+      "step": 740
+    },
+    {
+      "epoch": 0.8116883116883117,
+      "grad_norm": 0.185546875,
+      "learning_rate": 0.00037836826347305394,
+      "loss": 0.1537,
+      "step": 750
+    },
+    {
+      "epoch": 0.8225108225108225,
+      "grad_norm": 0.2021484375,
+      "learning_rate": 0.00037649700598802393,
+      "loss": 0.1617,
+      "step": 760
+    },
+    {
+      "epoch": 0.8333333333333334,
+      "grad_norm": 0.267578125,
+      "learning_rate": 0.00037462574850299403,
+      "loss": 0.1641,
+      "step": 770
+    },
+    {
+      "epoch": 0.8441558441558441,
+      "grad_norm": 0.296875,
+      "learning_rate": 0.0003727544910179641,
+      "loss": 0.1742,
+      "step": 780
+    },
+    {
+      "epoch": 0.854978354978355,
+      "grad_norm": 0.2265625,
+      "learning_rate": 0.0003708832335329341,
+      "loss": 0.1787,
+      "step": 790
+    },
+    {
+      "epoch": 0.8658008658008658,
+      "grad_norm": 0.1845703125,
+      "learning_rate": 0.0003690119760479042,
+      "loss": 0.1736,
+      "step": 800
+    },
+    {
+      "epoch": 0.8658008658008658,
+      "eval_loss": 0.16875208914279938,
+      "eval_runtime": 49.7255,
+      "eval_samples_per_second": 10.055,
+      "eval_steps_per_second": 0.08,
+      "step": 800
+    },
+    {
+      "epoch": 0.8766233766233766,
+      "grad_norm": 0.2109375,
+      "learning_rate": 0.00036714071856287425,
+      "loss": 0.1729,
+      "step": 810
+    },
+    {
+      "epoch": 0.8874458874458875,
+      "grad_norm": 0.392578125,
+      "learning_rate": 0.0003652694610778443,
+      "loss": 0.1617,
+      "step": 820
+    },
+    {
+      "epoch": 0.8982683982683982,
+      "grad_norm": 0.2578125,
+      "learning_rate": 0.0003633982035928144,
+      "loss": 0.1637,
+      "step": 830
+    },
+    {
+      "epoch": 0.9090909090909091,
+      "grad_norm": 0.2109375,
+      "learning_rate": 0.00036152694610778444,
+      "loss": 0.1569,
+      "step": 840
+    },
+    {
+      "epoch": 0.9199134199134199,
+      "grad_norm": 0.18359375,
+      "learning_rate": 0.0003596556886227545,
+      "loss": 0.1547,
+      "step": 850
+    },
+    {
+      "epoch": 0.9307359307359307,
+      "grad_norm": 0.2197265625,
+      "learning_rate": 0.0003577844311377246,
+      "loss": 0.1645,
+      "step": 860
+    },
+    {
+      "epoch": 0.9415584415584416,
+      "grad_norm": 0.357421875,
+      "learning_rate": 0.0003559131736526946,
+      "loss": 0.163,
+      "step": 870
+    },
+    {
+      "epoch": 0.9523809523809523,
+      "grad_norm": 0.255859375,
+      "learning_rate": 0.00035404191616766466,
+      "loss": 0.1748,
+      "step": 880
+    },
+    {
+      "epoch": 0.9632034632034632,
+      "grad_norm": 0.19921875,
+      "learning_rate": 0.00035217065868263476,
+      "loss": 0.1672,
+      "step": 890
+    },
+    {
+      "epoch": 0.974025974025974,
+      "grad_norm": 0.330078125,
+      "learning_rate": 0.0003502994011976048,
+      "loss": 0.1699,
+      "step": 900
+    },
+    {
+      "epoch": 0.974025974025974,
+      "eval_loss": 0.16794857382774353,
+      "eval_runtime": 58.792,
+      "eval_samples_per_second": 8.505,
+      "eval_steps_per_second": 0.068,
+      "step": 900
+    },
+    {
+      "epoch": 0.9848484848484849,
+      "grad_norm": 0.1435546875,
+      "learning_rate": 0.00034842814371257485,
+      "loss": 0.1537,
+      "step": 910
+    },
+    {
+      "epoch": 0.9956709956709957,
+      "grad_norm": 0.2421875,
+      "learning_rate": 0.00034655688622754494,
+      "loss": 0.1629,
+      "step": 920
+    },
+    {
+      "epoch": 1.0064935064935066,
+      "grad_norm": 0.19140625,
+      "learning_rate": 0.000344685628742515,
+      "loss": 0.1499,
+      "step": 930
+    },
+    {
+      "epoch": 1.0173160173160174,
+      "grad_norm": 0.28125,
+      "learning_rate": 0.00034281437125748503,
+      "loss": 0.1464,
+      "step": 940
+    },
+    {
+      "epoch": 1.0281385281385282,
+      "grad_norm": 0.173828125,
+      "learning_rate": 0.00034094311377245507,
+      "loss": 0.1416,
+      "step": 950
+    },
+    {
+      "epoch": 1.0389610389610389,
+      "grad_norm": 0.2119140625,
+      "learning_rate": 0.00033907185628742517,
+      "loss": 0.1545,
+      "step": 960
+    },
+    {
+      "epoch": 1.0497835497835497,
+      "grad_norm": 0.18359375,
+      "learning_rate": 0.0003372005988023952,
+      "loss": 0.1431,
+      "step": 970
+    },
+    {
+      "epoch": 1.0606060606060606,
+      "grad_norm": 0.279296875,
+      "learning_rate": 0.00033532934131736525,
+      "loss": 0.1607,
+      "step": 980
+    },
+    {
+      "epoch": 1.0714285714285714,
+      "grad_norm": 0.384765625,
+      "learning_rate": 0.00033345808383233535,
+      "loss": 0.1602,
+      "step": 990
+    },
+    {
+      "epoch": 1.0822510822510822,
+      "grad_norm": 0.2470703125,
+      "learning_rate": 0.0003315868263473054,
+      "loss": 0.1651,
+      "step": 1000
+    },
+    {
+      "epoch": 1.0822510822510822,
+      "eval_loss": 0.1569806933403015,
+      "eval_runtime": 57.763,
+      "eval_samples_per_second": 8.656,
+      "eval_steps_per_second": 0.069,
+      "step": 1000
+    },
+    {
+      "epoch": 1.093073593073593,
+      "grad_norm": 0.2734375,
+      "learning_rate": 0.00032971556886227544,
+      "loss": 0.146,
+      "step": 1010
+    },
+    {
+      "epoch": 1.103896103896104,
+      "grad_norm": 0.275390625,
+      "learning_rate": 0.00032784431137724553,
+      "loss": 0.1341,
+      "step": 1020
+    },
+    {
+      "epoch": 1.1147186147186148,
+      "grad_norm": 0.1640625,
+      "learning_rate": 0.0003259730538922156,
+      "loss": 0.1412,
+      "step": 1030
+    },
+    {
+      "epoch": 1.1255411255411256,
+      "grad_norm": 0.21875,
+      "learning_rate": 0.0003241017964071856,
+      "loss": 0.152,
+      "step": 1040
+    },
+    {
+      "epoch": 1.1363636363636362,
+      "grad_norm": 0.1962890625,
+      "learning_rate": 0.0003222305389221557,
+      "loss": 0.1501,
+      "step": 1050
+    },
+    {
+      "epoch": 1.1471861471861473,
+      "grad_norm": 0.1337890625,
+      "learning_rate": 0.00032035928143712576,
+      "loss": 0.1359,
+      "step": 1060
+    },
+    {
+      "epoch": 1.158008658008658,
+      "grad_norm": 0.271484375,
+      "learning_rate": 0.0003184880239520958,
+      "loss": 0.1437,
+      "step": 1070
+    },
+    {
+      "epoch": 1.1688311688311688,
+      "grad_norm": 0.2216796875,
+      "learning_rate": 0.0003166167664670659,
+      "loss": 0.1395,
+      "step": 1080
+    },
+    {
+      "epoch": 1.1796536796536796,
+      "grad_norm": 0.314453125,
+      "learning_rate": 0.0003147455089820359,
+      "loss": 0.1403,
+      "step": 1090
+    },
+    {
+      "epoch": 1.1904761904761905,
+      "grad_norm": 0.2236328125,
+      "learning_rate": 0.000312874251497006,
+      "loss": 0.157,
+      "step": 1100
+    },
+    {
+      "epoch": 1.1904761904761905,
+      "eval_loss": 0.15362851321697235,
+      "eval_runtime": 50.1243,
+      "eval_samples_per_second": 9.975,
+      "eval_steps_per_second": 0.08,
+      "step": 1100
+    },
+    {
+      "epoch": 1.2012987012987013,
+      "grad_norm": 0.2177734375,
+      "learning_rate": 0.0003110029940119761,
+      "loss": 0.1377,
+      "step": 1110
+    },
+    {
+      "epoch": 1.2121212121212122,
+      "grad_norm": 0.2138671875,
+      "learning_rate": 0.0003091317365269461,
+      "loss": 0.1423,
+      "step": 1120
+    },
+    {
+      "epoch": 1.222943722943723,
+      "grad_norm": 0.1953125,
+      "learning_rate": 0.00030726047904191617,
+      "loss": 0.1353,
+      "step": 1130
+    },
+    {
+      "epoch": 1.2337662337662338,
+      "grad_norm": 0.255859375,
+      "learning_rate": 0.00030538922155688627,
+      "loss": 0.1453,
+      "step": 1140
+    },
+    {
+      "epoch": 1.2445887445887447,
+      "grad_norm": 0.3828125,
+      "learning_rate": 0.00030351796407185626,
+      "loss": 0.1398,
+      "step": 1150
+    },
+    {
+      "epoch": 1.2554112554112553,
+      "grad_norm": 0.189453125,
+      "learning_rate": 0.00030164670658682635,
+      "loss": 0.1315,
+      "step": 1160
+    },
+    {
+      "epoch": 1.2662337662337662,
+      "grad_norm": 0.26953125,
+      "learning_rate": 0.00029977544910179645,
+      "loss": 0.1323,
+      "step": 1170
+    },
+    {
+      "epoch": 1.277056277056277,
+      "grad_norm": 0.1513671875,
+      "learning_rate": 0.00029790419161676644,
+      "loss": 0.134,
+      "step": 1180
+    },
+    {
+      "epoch": 1.2878787878787878,
+      "grad_norm": 0.3046875,
+      "learning_rate": 0.00029603293413173654,
+      "loss": 0.1544,
+      "step": 1190
+    },
+    {
+      "epoch": 1.2987012987012987,
+      "grad_norm": 0.236328125,
+      "learning_rate": 0.0002941616766467066,
+      "loss": 0.141,
+      "step": 1200
+    },
+    {
+      "epoch": 1.2987012987012987,
+      "eval_loss": 0.1604161411523819,
+      "eval_runtime": 51.9642,
+      "eval_samples_per_second": 9.622,
+      "eval_steps_per_second": 0.077,
+      "step": 1200
+    },
+    {
+      "epoch": 1.3095238095238095,
+      "grad_norm": 0.14453125,
+      "learning_rate": 0.0002922904191616766,
+      "loss": 0.1405,
+      "step": 1210
+    },
+    {
+      "epoch": 1.3203463203463204,
+      "grad_norm": 0.154296875,
+      "learning_rate": 0.0002904191616766467,
+      "loss": 0.1388,
+      "step": 1220
+    },
+    {
+      "epoch": 1.3311688311688312,
+      "grad_norm": 0.1572265625,
+      "learning_rate": 0.00028854790419161676,
+      "loss": 0.1482,
+      "step": 1230
+    },
+    {
+      "epoch": 1.341991341991342,
+      "grad_norm": 0.201171875,
+      "learning_rate": 0.0002866766467065868,
+      "loss": 0.1299,
+      "step": 1240
+    },
+    {
+      "epoch": 1.3528138528138527,
+      "grad_norm": 0.2265625,
+      "learning_rate": 0.0002848053892215569,
+      "loss": 0.137,
+      "step": 1250
+    },
+    {
+      "epoch": 1.3636363636363638,
+      "grad_norm": 0.2255859375,
+      "learning_rate": 0.00028293413173652695,
+      "loss": 0.1173,
+      "step": 1260
+    },
+    {
+      "epoch": 1.3744588744588744,
+      "grad_norm": 0.25,
+      "learning_rate": 0.000281062874251497,
+      "loss": 0.1317,
+      "step": 1270
+    },
+    {
+      "epoch": 1.3852813852813852,
+      "grad_norm": 0.1455078125,
+      "learning_rate": 0.0002791916167664671,
+      "loss": 0.1475,
+      "step": 1280
+    },
+    {
+      "epoch": 1.396103896103896,
+      "grad_norm": 0.150390625,
+      "learning_rate": 0.00027732035928143713,
+      "loss": 0.1411,
+      "step": 1290
+    },
+    {
+      "epoch": 1.406926406926407,
+      "grad_norm": 0.1552734375,
+      "learning_rate": 0.00027544910179640717,
+      "loss": 0.1343,
+      "step": 1300
+    },
+    {
+      "epoch": 1.406926406926407,
+      "eval_loss": 0.1617971956729889,
+      "eval_runtime": 53.8477,
+      "eval_samples_per_second": 9.285,
+      "eval_steps_per_second": 0.074,
+      "step": 1300
+    },
+    {
+      "epoch": 1.4177489177489178,
+      "grad_norm": 0.26171875,
+      "learning_rate": 0.00027357784431137727,
+      "loss": 0.1367,
+      "step": 1310
+    },
+    {
+      "epoch": 1.4285714285714286,
+      "grad_norm": 0.154296875,
+      "learning_rate": 0.0002717065868263473,
+      "loss": 0.1359,
+      "step": 1320
+    },
+    {
+      "epoch": 1.4393939393939394,
+      "grad_norm": 0.208984375,
+      "learning_rate": 0.00026983532934131735,
+      "loss": 0.1257,
+      "step": 1330
+    },
+    {
+      "epoch": 1.4502164502164503,
+      "grad_norm": 0.1865234375,
+      "learning_rate": 0.0002679640718562874,
+      "loss": 0.1282,
+      "step": 1340
+    },
+    {
+      "epoch": 1.4610389610389611,
+      "grad_norm": 0.1767578125,
+      "learning_rate": 0.0002660928143712575,
+      "loss": 0.1266,
+      "step": 1350
+    },
+    {
+      "epoch": 1.4718614718614718,
+      "grad_norm": 0.1982421875,
+      "learning_rate": 0.0002642215568862276,
+      "loss": 0.1204,
+      "step": 1360
+    },
+    {
+      "epoch": 1.4826839826839826,
+      "grad_norm": 0.1962890625,
+      "learning_rate": 0.0002623502994011976,
+      "loss": 0.151,
+      "step": 1370
+    },
+    {
+      "epoch": 1.4935064935064934,
+      "grad_norm": 0.20703125,
+      "learning_rate": 0.0002604790419161677,
+      "loss": 0.1336,
+      "step": 1380
+    },
+    {
+      "epoch": 1.5043290043290043,
+      "grad_norm": 0.1376953125,
+      "learning_rate": 0.0002586077844311378,
+      "loss": 0.1364,
+      "step": 1390
+    },
+    {
+      "epoch": 1.5151515151515151,
+      "grad_norm": 0.333984375,
+      "learning_rate": 0.00025673652694610776,
+      "loss": 0.1466,
+      "step": 1400
+    },
+    {
+      "epoch": 1.5151515151515151,
+      "eval_loss": 0.165997713804245,
+      "eval_runtime": 49.4239,
+      "eval_samples_per_second": 10.117,
+      "eval_steps_per_second": 0.081,
+      "step": 1400
+    },
+    {
+      "epoch": 1.525974025974026,
+      "grad_norm": 0.154296875,
+      "learning_rate": 0.00025486526946107786,
+      "loss": 0.1306,
+      "step": 1410
+    },
+    {
+      "epoch": 1.5367965367965368,
+      "grad_norm": 0.1630859375,
+      "learning_rate": 0.00025299401197604796,
+      "loss": 0.1246,
+      "step": 1420
+    },
+    {
+      "epoch": 1.5476190476190477,
+      "grad_norm": 0.1904296875,
+      "learning_rate": 0.00025112275449101795,
+      "loss": 0.1333,
+      "step": 1430
+    },
+    {
+      "epoch": 1.5584415584415585,
+      "grad_norm": 0.201171875,
+      "learning_rate": 0.00024925149700598804,
+      "loss": 0.1258,
+      "step": 1440
+    },
+    {
+      "epoch": 1.5692640692640691,
+      "grad_norm": 0.244140625,
+      "learning_rate": 0.0002473802395209581,
+      "loss": 0.1493,
+      "step": 1450
+    },
+    {
+      "epoch": 1.5800865800865802,
+      "grad_norm": 0.1279296875,
+      "learning_rate": 0.00024550898203592813,
+      "loss": 0.1346,
+      "step": 1460
+    },
+    {
+      "epoch": 1.5909090909090908,
+      "grad_norm": 0.154296875,
+      "learning_rate": 0.00024363772455089823,
+      "loss": 0.1228,
+      "step": 1470
+    },
+    {
+      "epoch": 1.601731601731602,
+      "grad_norm": 0.22265625,
+      "learning_rate": 0.00024176646706586827,
+      "loss": 0.1335,
+      "step": 1480
+    },
+    {
+      "epoch": 1.6125541125541125,
+      "grad_norm": 0.2333984375,
+      "learning_rate": 0.0002398952095808383,
+      "loss": 0.137,
+      "step": 1490
+    },
+    {
+      "epoch": 1.6233766233766234,
+      "grad_norm": 0.197265625,
+      "learning_rate": 0.00023802395209580838,
+      "loss": 0.1351,
+      "step": 1500
+    },
+    {
+      "epoch": 1.6233766233766234,
+      "eval_loss": 0.1522991955280304,
+      "eval_runtime": 49.2679,
+      "eval_samples_per_second": 10.149,
+      "eval_steps_per_second": 0.081,
+      "step": 1500
+    },
+    {
+      "epoch": 1.6341991341991342,
+      "grad_norm": 0.330078125,
+      "learning_rate": 0.00023615269461077845,
+      "loss": 0.1361,
+      "step": 1510
+    },
+    {
+      "epoch": 1.645021645021645,
+      "grad_norm": 0.140625,
+      "learning_rate": 0.0002342814371257485,
+      "loss": 0.1383,
+      "step": 1520
+    },
+    {
+      "epoch": 1.655844155844156,
+      "grad_norm": 0.1826171875,
+      "learning_rate": 0.00023241017964071857,
+      "loss": 0.135,
+      "step": 1530
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 0.2197265625,
+      "learning_rate": 0.00023053892215568864,
+      "loss": 0.1298,
+      "step": 1540
+    },
+    {
+      "epoch": 1.6774891774891776,
+      "grad_norm": 0.2080078125,
+      "learning_rate": 0.00022866766467065868,
+      "loss": 0.1323,
+      "step": 1550
+    },
+    {
+      "epoch": 1.6883116883116882,
+      "grad_norm": 0.232421875,
+      "learning_rate": 0.00022679640718562875,
+      "loss": 0.1253,
+      "step": 1560
+    },
+    {
+      "epoch": 1.6991341991341993,
+      "grad_norm": 0.130859375,
+      "learning_rate": 0.0002249251497005988,
+      "loss": 0.1324,
+      "step": 1570
+    },
+    {
+      "epoch": 1.70995670995671,
+      "grad_norm": 0.1865234375,
+      "learning_rate": 0.0002230538922155689,
+      "loss": 0.1402,
+      "step": 1580
+    },
+    {
+      "epoch": 1.7207792207792207,
+      "grad_norm": 0.1875,
+      "learning_rate": 0.00022118263473053893,
+      "loss": 0.1307,
+      "step": 1590
+    },
+    {
+      "epoch": 1.7316017316017316,
+      "grad_norm": 0.12890625,
+      "learning_rate": 0.00021931137724550898,
+      "loss": 0.1276,
+      "step": 1600
+    },
+    {
+      "epoch": 1.7316017316017316,
+      "eval_loss": 0.1606496274471283,
+      "eval_runtime": 51.8902,
+      "eval_samples_per_second": 9.636,
+      "eval_steps_per_second": 0.077,
+      "step": 1600
+    },
+    {
+      "epoch": 1.7424242424242424,
+      "grad_norm": 0.1455078125,
+      "learning_rate": 0.00021744011976047905,
+      "loss": 0.1302,
+      "step": 1610
+    },
+    {
+      "epoch": 1.7532467532467533,
+      "grad_norm": 0.1455078125,
+      "learning_rate": 0.00021556886227544912,
+      "loss": 0.1231,
+      "step": 1620
+    },
+    {
+      "epoch": 1.7640692640692641,
+      "grad_norm": 0.162109375,
+      "learning_rate": 0.00021369760479041916,
+      "loss": 0.1362,
+      "step": 1630
+    },
+    {
+      "epoch": 1.774891774891775,
+      "grad_norm": 0.16796875,
+      "learning_rate": 0.00021182634730538923,
+      "loss": 0.1264,
+      "step": 1640
+    },
+    {
+      "epoch": 1.7857142857142856,
+      "grad_norm": 0.25390625,
+      "learning_rate": 0.0002099550898203593,
+      "loss": 0.142,
+      "step": 1650
+    },
+    {
+      "epoch": 1.7965367965367967,
+      "grad_norm": 0.1279296875,
+      "learning_rate": 0.00020808383233532934,
+      "loss": 0.1347,
+      "step": 1660
+    },
+    {
+      "epoch": 1.8073593073593073,
+      "grad_norm": 0.1708984375,
+      "learning_rate": 0.0002062125748502994,
+      "loss": 0.1399,
+      "step": 1670
+    },
+    {
+      "epoch": 1.8181818181818183,
+      "grad_norm": 0.24609375,
+      "learning_rate": 0.00020434131736526945,
+      "loss": 0.1236,
+      "step": 1680
+    },
+    {
+      "epoch": 1.829004329004329,
+      "grad_norm": 0.2255859375,
+      "learning_rate": 0.00020247005988023952,
+      "loss": 0.113,
+      "step": 1690
+    },
+    {
+      "epoch": 1.8398268398268398,
+      "grad_norm": 0.2421875,
+      "learning_rate": 0.0002005988023952096,
+      "loss": 0.1239,
+      "step": 1700
+    },
+    {
+      "epoch": 1.8398268398268398,
+      "eval_loss": 0.16243509948253632,
+      "eval_runtime": 54.1441,
+      "eval_samples_per_second": 9.235,
+      "eval_steps_per_second": 0.074,
+      "step": 1700
+    },
+    {
+      "epoch": 1.8506493506493507,
+      "grad_norm": 0.1455078125,
+      "learning_rate": 0.00019872754491017964,
+      "loss": 0.1257,
+      "step": 1710
+    },
+    {
+      "epoch": 1.8614718614718615,
+      "grad_norm": 0.1982421875,
+      "learning_rate": 0.0001968562874251497,
+      "loss": 0.151,
+      "step": 1720
+    },
+    {
+      "epoch": 1.8722943722943723,
+      "grad_norm": 0.30078125,
+      "learning_rate": 0.00019498502994011978,
+      "loss": 0.1238,
+      "step": 1730
+    },
+    {
+      "epoch": 1.883116883116883,
+      "grad_norm": 0.2734375,
+      "learning_rate": 0.00019311377245508982,
+      "loss": 0.1447,
+      "step": 1740
+    },
+    {
+      "epoch": 1.893939393939394,
+      "grad_norm": 0.1728515625,
+      "learning_rate": 0.00019124251497005986,
+      "loss": 0.14,
+      "step": 1750
+    },
+    {
+      "epoch": 1.9047619047619047,
+      "grad_norm": 0.146484375,
+      "learning_rate": 0.00018937125748502996,
+      "loss": 0.118,
+      "step": 1760
+    },
+    {
+      "epoch": 1.9155844155844157,
+      "grad_norm": 0.275390625,
+      "learning_rate": 0.0001875,
+      "loss": 0.126,
+      "step": 1770
+    },
+    {
+      "epoch": 1.9264069264069263,
+      "grad_norm": 0.19140625,
+      "learning_rate": 0.00018562874251497005,
+      "loss": 0.1399,
+      "step": 1780
+    },
+    {
+      "epoch": 1.9372294372294372,
+      "grad_norm": 0.240234375,
+      "learning_rate": 0.00018375748502994014,
+      "loss": 0.1291,
+      "step": 1790
+    },
+    {
+      "epoch": 1.948051948051948,
+      "grad_norm": 0.205078125,
+      "learning_rate": 0.0001818862275449102,
+      "loss": 0.1339,
+      "step": 1800
+    },
+    {
+      "epoch": 1.948051948051948,
+      "eval_loss": 0.1591983437538147,
+      "eval_runtime": 47.2348,
+      "eval_samples_per_second": 10.585,
+      "eval_steps_per_second": 0.085,
+      "step": 1800
+    },
+    {
+      "epoch": 1.9588744588744589,
+      "grad_norm": 0.22265625,
+      "learning_rate": 0.00018001497005988023,
+      "loss": 0.1414,
+      "step": 1810
+    },
+    {
+      "epoch": 1.9696969696969697,
+      "grad_norm": 0.2080078125,
+      "learning_rate": 0.0001781437125748503,
+      "loss": 0.116,
+      "step": 1820
+    },
+    {
+      "epoch": 1.9805194805194806,
+      "grad_norm": 0.265625,
+      "learning_rate": 0.00017627245508982037,
+      "loss": 0.1279,
+      "step": 1830
+    },
+    {
+      "epoch": 1.9913419913419914,
+      "grad_norm": 0.2236328125,
+      "learning_rate": 0.0001744011976047904,
+      "loss": 0.1136,
+      "step": 1840
+    },
+    {
+      "epoch": 2.002164502164502,
+      "grad_norm": 0.11669921875,
+      "learning_rate": 0.00017252994011976048,
+      "loss": 0.1158,
+      "step": 1850
+    },
+    {
+      "epoch": 2.012987012987013,
+      "grad_norm": 0.12353515625,
+      "learning_rate": 0.00017065868263473055,
+      "loss": 0.113,
+      "step": 1860
+    },
+    {
+      "epoch": 2.0238095238095237,
+      "grad_norm": 0.234375,
+      "learning_rate": 0.0001687874251497006,
+      "loss": 0.1127,
+      "step": 1870
+    },
+    {
+      "epoch": 2.034632034632035,
+      "grad_norm": 0.1728515625,
+      "learning_rate": 0.00016691616766467067,
+      "loss": 0.1289,
+      "step": 1880
+    },
+    {
+      "epoch": 2.0454545454545454,
+      "grad_norm": 0.10791015625,
+      "learning_rate": 0.0001650449101796407,
+      "loss": 0.1236,
+      "step": 1890
+    },
+    {
+      "epoch": 2.0562770562770565,
+      "grad_norm": 0.2080078125,
+      "learning_rate": 0.00016317365269461078,
+      "loss": 0.114,
+      "step": 1900
+    },
+    {
+      "epoch": 2.0562770562770565,
+      "eval_loss": 0.15889017283916473,
+      "eval_runtime": 49.4787,
+      "eval_samples_per_second": 10.105,
+      "eval_steps_per_second": 0.081,
+      "step": 1900
+    },
+    {
+      "epoch": 2.067099567099567,
+      "grad_norm": 0.1875,
+      "learning_rate": 0.00016130239520958085,
+      "loss": 0.1202,
+      "step": 1910
+    },
+    {
+      "epoch": 2.0779220779220777,
+      "grad_norm": 0.30078125,
+      "learning_rate": 0.0001594311377245509,
+      "loss": 0.1236,
+      "step": 1920
+    },
+    {
+      "epoch": 2.088744588744589,
+      "grad_norm": 0.1630859375,
+      "learning_rate": 0.00015755988023952096,
+      "loss": 0.1151,
+      "step": 1930
+    },
+    {
+      "epoch": 2.0995670995670994,
+      "grad_norm": 0.1806640625,
+      "learning_rate": 0.00015568862275449103,
+      "loss": 0.116,
+      "step": 1940
+    },
+    {
+      "epoch": 2.1103896103896105,
+      "grad_norm": 0.20703125,
+      "learning_rate": 0.00015381736526946108,
+      "loss": 0.1376,
+      "step": 1950
+    },
+    {
+      "epoch": 2.121212121212121,
+      "grad_norm": 0.1865234375,
+      "learning_rate": 0.00015194610778443112,
+      "loss": 0.1189,
+      "step": 1960
+    },
+    {
+      "epoch": 2.132034632034632,
+      "grad_norm": 0.1103515625,
+      "learning_rate": 0.00015007485029940122,
+      "loss": 0.1122,
+      "step": 1970
+    },
+    {
+      "epoch": 2.142857142857143,
+      "grad_norm": 0.1494140625,
+      "learning_rate": 0.00014820359281437126,
+      "loss": 0.1159,
+      "step": 1980
+    },
+    {
+      "epoch": 2.153679653679654,
+      "grad_norm": 0.1826171875,
+      "learning_rate": 0.00014633233532934133,
+      "loss": 0.1139,
+      "step": 1990
+    },
+    {
+      "epoch": 2.1645021645021645,
+      "grad_norm": 0.234375,
+      "learning_rate": 0.00014446107784431137,
+      "loss": 0.1173,
+      "step": 2000
+    },
+    {
+      "epoch": 2.1645021645021645,
+      "eval_loss": 0.16071972250938416,
+      "eval_runtime": 53.7032,
+      "eval_samples_per_second": 9.31,
+      "eval_steps_per_second": 0.074,
+      "step": 2000
+    },
+    {
+      "epoch": 2.175324675324675,
+      "grad_norm": 0.10498046875,
+      "learning_rate": 0.00014258982035928144,
+      "loss": 0.0994,
+      "step": 2010
+    },
+    {
+      "epoch": 2.186147186147186,
+      "grad_norm": 0.1630859375,
+      "learning_rate": 0.0001407185628742515,
+      "loss": 0.1135,
+      "step": 2020
+    },
+    {
+      "epoch": 2.196969696969697,
+      "grad_norm": 0.1884765625,
+      "learning_rate": 0.00013884730538922155,
+      "loss": 0.1186,
+      "step": 2030
+    },
+    {
+      "epoch": 2.207792207792208,
+      "grad_norm": 0.1904296875,
+      "learning_rate": 0.00013697604790419162,
+      "loss": 0.1066,
+      "step": 2040
+    },
+    {
+      "epoch": 2.2186147186147185,
+      "grad_norm": 0.1923828125,
+      "learning_rate": 0.0001351047904191617,
+      "loss": 0.0979,
+      "step": 2050
+    },
+    {
+      "epoch": 2.2294372294372296,
+      "grad_norm": 0.10546875,
+      "learning_rate": 0.00013323353293413174,
+      "loss": 0.1214,
+      "step": 2060
+    },
+    {
+      "epoch": 2.24025974025974,
+      "grad_norm": 0.1259765625,
+      "learning_rate": 0.00013136227544910178,
+      "loss": 0.1086,
+      "step": 2070
+    },
+    {
+      "epoch": 2.2510822510822512,
+      "grad_norm": 0.2060546875,
+      "learning_rate": 0.00012949101796407188,
+      "loss": 0.0932,
+      "step": 2080
+    },
+    {
+      "epoch": 2.261904761904762,
+      "grad_norm": 0.1318359375,
+      "learning_rate": 0.00012761976047904192,
+      "loss": 0.1053,
+      "step": 2090
+    },
+    {
+      "epoch": 2.2727272727272725,
+      "grad_norm": 0.16015625,
+      "learning_rate": 0.00012574850299401196,
+      "loss": 0.1231,
+      "step": 2100
+    },
+    {
+      "epoch": 2.2727272727272725,
+      "eval_loss": 0.1609112173318863,
+      "eval_runtime": 50.2755,
+      "eval_samples_per_second": 9.945,
+      "eval_steps_per_second": 0.08,
+      "step": 2100
+    },
+    {
+      "epoch": 2.2835497835497836,
+      "grad_norm": 0.1357421875,
+      "learning_rate": 0.00012387724550898203,
+      "loss": 0.1207,
+      "step": 2110
+    },
+    {
+      "epoch": 2.2943722943722946,
+      "grad_norm": 0.16015625,
+      "learning_rate": 0.0001220059880239521,
+      "loss": 0.1202,
+      "step": 2120
+    },
+    {
+      "epoch": 2.3051948051948052,
+      "grad_norm": 0.1943359375,
+      "learning_rate": 0.00012013473053892216,
+      "loss": 0.1215,
+      "step": 2130
+    },
+    {
+      "epoch": 2.316017316017316,
+      "grad_norm": 0.2314453125,
+      "learning_rate": 0.00011826347305389222,
+      "loss": 0.112,
+      "step": 2140
+    },
+    {
+      "epoch": 2.326839826839827,
+      "grad_norm": 0.1767578125,
+      "learning_rate": 0.00011639221556886227,
+      "loss": 0.1233,
+      "step": 2150
+    },
+    {
+      "epoch": 2.3376623376623376,
+      "grad_norm": 0.16015625,
+      "learning_rate": 0.00011452095808383234,
+      "loss": 0.1086,
+      "step": 2160
+    },
+    {
+      "epoch": 2.3484848484848486,
+      "grad_norm": 0.138671875,
+      "learning_rate": 0.0001126497005988024,
+      "loss": 0.1123,
+      "step": 2170
+    },
+    {
+      "epoch": 2.3593073593073592,
+      "grad_norm": 0.193359375,
+      "learning_rate": 0.00011077844311377246,
+      "loss": 0.1145,
+      "step": 2180
+    },
+    {
+      "epoch": 2.3701298701298703,
+      "grad_norm": 0.2216796875,
+      "learning_rate": 0.00010890718562874253,
+      "loss": 0.1117,
+      "step": 2190
+    },
+    {
+      "epoch": 2.380952380952381,
+      "grad_norm": 0.2890625,
+      "learning_rate": 0.00010703592814371257,
+      "loss": 0.12,
+      "step": 2200
+    },
+    {
+      "epoch": 2.380952380952381,
+      "eval_loss": 0.16090266406536102,
+      "eval_runtime": 59.1415,
+      "eval_samples_per_second": 8.454,
+      "eval_steps_per_second": 0.068,
+      "step": 2200
+    },
+    {
+      "epoch": 2.391774891774892,
+      "grad_norm": 0.220703125,
+      "learning_rate": 0.00010516467065868264,
+      "loss": 0.11,
+      "step": 2210
+    },
+    {
+      "epoch": 2.4025974025974026,
+      "grad_norm": 0.1728515625,
+      "learning_rate": 0.0001032934131736527,
+      "loss": 0.1198,
+      "step": 2220
+    },
+    {
+      "epoch": 2.4134199134199132,
+      "grad_norm": 0.23828125,
+      "learning_rate": 0.00010142215568862275,
+      "loss": 0.1179,
+      "step": 2230
+    },
+    {
+      "epoch": 2.4242424242424243,
+      "grad_norm": 0.197265625,
+      "learning_rate": 9.955089820359281e-05,
+      "loss": 0.1146,
+      "step": 2240
+    },
+    {
+      "epoch": 2.435064935064935,
+      "grad_norm": 0.1953125,
+      "learning_rate": 9.767964071856288e-05,
+      "loss": 0.1067,
+      "step": 2250
+    },
+    {
+      "epoch": 2.445887445887446,
+      "grad_norm": 0.158203125,
+      "learning_rate": 9.580838323353294e-05,
+      "loss": 0.105,
+      "step": 2260
+    },
+    {
+      "epoch": 2.4567099567099566,
+      "grad_norm": 0.189453125,
+      "learning_rate": 9.393712574850299e-05,
+      "loss": 0.1222,
+      "step": 2270
+    },
+    {
+      "epoch": 2.4675324675324677,
+      "grad_norm": 0.1376953125,
+      "learning_rate": 9.206586826347306e-05,
+      "loss": 0.1068,
+      "step": 2280
+    },
+    {
+      "epoch": 2.4783549783549783,
+      "grad_norm": 0.1640625,
+      "learning_rate": 9.01946107784431e-05,
+      "loss": 0.1069,
+      "step": 2290
+    },
+    {
+      "epoch": 2.4891774891774894,
+      "grad_norm": 0.1728515625,
+      "learning_rate": 8.832335329341318e-05,
+      "loss": 0.1276,
+      "step": 2300
+    },
+    {
+      "epoch": 2.4891774891774894,
+      "eval_loss": 0.15976013243198395,
+      "eval_runtime": 51.0014,
+      "eval_samples_per_second": 9.804,
+      "eval_steps_per_second": 0.078,
+      "step": 2300
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 0.2431640625,
+      "learning_rate": 8.645209580838323e-05,
+      "loss": 0.1248,
+      "step": 2310
+    },
+    {
+      "epoch": 2.5108225108225106,
+      "grad_norm": 0.146484375,
+      "learning_rate": 8.45808383233533e-05,
+      "loss": 0.1062,
+      "step": 2320
+    },
+    {
+      "epoch": 2.5216450216450217,
+      "grad_norm": 0.1279296875,
+      "learning_rate": 8.270958083832336e-05,
+      "loss": 0.0997,
+      "step": 2330
+    },
+    {
+      "epoch": 2.5324675324675323,
+      "grad_norm": 0.1962890625,
+      "learning_rate": 8.083832335329341e-05,
+      "loss": 0.1171,
+      "step": 2340
+    },
+    {
+      "epoch": 2.5432900432900434,
+      "grad_norm": 0.19921875,
+      "learning_rate": 7.896706586826349e-05,
+      "loss": 0.0944,
+      "step": 2350
+    },
+    {
+      "epoch": 2.554112554112554,
+      "grad_norm": 0.22265625,
+      "learning_rate": 7.709580838323353e-05,
+      "loss": 0.1091,
+      "step": 2360
+    },
+    {
+      "epoch": 2.564935064935065,
+      "grad_norm": 0.2099609375,
+      "learning_rate": 7.52245508982036e-05,
+      "loss": 0.1151,
+      "step": 2370
+    },
+    {
+      "epoch": 2.5757575757575757,
+      "grad_norm": 0.1142578125,
+      "learning_rate": 7.335329341317365e-05,
+      "loss": 0.1174,
+      "step": 2380
+    },
+    {
+      "epoch": 2.5865800865800868,
+      "grad_norm": 0.1494140625,
+      "learning_rate": 7.148203592814371e-05,
+      "loss": 0.1219,
+      "step": 2390
+    },
+    {
+      "epoch": 2.5974025974025974,
+      "grad_norm": 0.244140625,
+      "learning_rate": 6.961077844311377e-05,
+      "loss": 0.1251,
+      "step": 2400
+    },
+    {
+      "epoch": 2.5974025974025974,
+      "eval_loss": 0.1569649577140808,
+      "eval_runtime": 50.4125,
+      "eval_samples_per_second": 9.918,
+      "eval_steps_per_second": 0.079,
+      "step": 2400
+    },
+    {
+      "epoch": 2.608225108225108,
+      "grad_norm": 0.34375,
+      "learning_rate": 6.773952095808384e-05,
+      "loss": 0.1171,
+      "step": 2410
+    },
+    {
+      "epoch": 2.619047619047619,
+      "grad_norm": 0.1630859375,
+      "learning_rate": 6.58682634730539e-05,
+      "loss": 0.1114,
+      "step": 2420
+    },
+    {
+      "epoch": 2.62987012987013,
+      "grad_norm": 0.208984375,
+      "learning_rate": 6.399700598802395e-05,
+      "loss": 0.1046,
+      "step": 2430
+    },
+    {
+      "epoch": 2.6406926406926408,
+      "grad_norm": 0.08154296875,
+      "learning_rate": 6.212574850299401e-05,
+      "loss": 0.1081,
+      "step": 2440
+    },
+    {
+      "epoch": 2.6515151515151514,
+      "grad_norm": 0.142578125,
+      "learning_rate": 6.025449101796408e-05,
+      "loss": 0.1042,
+      "step": 2450
+    },
+    {
+      "epoch": 2.6623376623376624,
+      "grad_norm": 0.1474609375,
+      "learning_rate": 5.8383233532934134e-05,
+      "loss": 0.0913,
+      "step": 2460
+    },
+    {
+      "epoch": 2.673160173160173,
+      "grad_norm": 0.330078125,
+      "learning_rate": 5.65119760479042e-05,
+      "loss": 0.1082,
+      "step": 2470
+    },
+    {
+      "epoch": 2.683982683982684,
+      "grad_norm": 0.173828125,
+      "learning_rate": 5.4640718562874254e-05,
+      "loss": 0.1158,
+      "step": 2480
+    },
+    {
+      "epoch": 2.6948051948051948,
+      "grad_norm": 0.1689453125,
+      "learning_rate": 5.276946107784431e-05,
+      "loss": 0.1135,
+      "step": 2490
+    },
+    {
+      "epoch": 2.7056277056277054,
+      "grad_norm": 0.2099609375,
+      "learning_rate": 5.089820359281437e-05,
+      "loss": 0.1332,
+      "step": 2500
+    },
+    {
+      "epoch": 2.7056277056277054,
+      "eval_loss": 0.15841823816299438,
+      "eval_runtime": 48.0255,
+      "eval_samples_per_second": 10.411,
+      "eval_steps_per_second": 0.083,
+      "step": 2500
+    },
+    {
+      "epoch": 2.7164502164502164,
+      "grad_norm": 0.140625,
+      "learning_rate": 4.902694610778443e-05,
+      "loss": 0.1076,
+      "step": 2510
+    },
+    {
+      "epoch": 2.7272727272727275,
+      "grad_norm": 0.11572265625,
+      "learning_rate": 4.7155688622754486e-05,
+      "loss": 0.1089,
+      "step": 2520
+    },
+    {
+      "epoch": 2.738095238095238,
+      "grad_norm": 0.1484375,
+      "learning_rate": 4.5284431137724556e-05,
+      "loss": 0.1061,
+      "step": 2530
+    },
+    {
+      "epoch": 2.7489177489177488,
+      "grad_norm": 0.123046875,
+      "learning_rate": 4.341317365269461e-05,
+      "loss": 0.1067,
+      "step": 2540
+    },
+    {
+      "epoch": 2.75974025974026,
+      "grad_norm": 0.201171875,
+      "learning_rate": 4.154191616766467e-05,
+      "loss": 0.1113,
+      "step": 2550
+    },
+    {
+      "epoch": 2.7705627705627704,
+      "grad_norm": 0.1494140625,
+      "learning_rate": 3.967065868263473e-05,
+      "loss": 0.1056,
+      "step": 2560
+    },
+    {
+      "epoch": 2.7813852813852815,
+      "grad_norm": 0.181640625,
+      "learning_rate": 3.779940119760479e-05,
+      "loss": 0.1135,
+      "step": 2570
+    },
+    {
+      "epoch": 2.792207792207792,
+      "grad_norm": 0.15625,
+      "learning_rate": 3.592814371257485e-05,
+      "loss": 0.1116,
+      "step": 2580
+    },
+    {
+      "epoch": 2.8030303030303028,
+      "grad_norm": 0.2255859375,
+      "learning_rate": 3.405688622754491e-05,
+      "loss": 0.1169,
+      "step": 2590
+    },
+    {
+      "epoch": 2.813852813852814,
+      "grad_norm": 0.08154296875,
+      "learning_rate": 3.2185628742514966e-05,
+      "loss": 0.1023,
+      "step": 2600
+    },
+    {
+      "epoch": 2.813852813852814,
+      "eval_loss": 0.15775570273399353,
+      "eval_runtime": 49.2082,
+      "eval_samples_per_second": 10.161,
+      "eval_steps_per_second": 0.081,
+      "step": 2600
+    },
+    {
+      "epoch": 2.824675324675325,
+      "grad_norm": 0.103515625,
+      "learning_rate": 3.031437125748503e-05,
+      "loss": 0.1076,
+      "step": 2610
+    },
+    {
+      "epoch": 2.8354978354978355,
+      "grad_norm": 0.1611328125,
+      "learning_rate": 2.844311377245509e-05,
+      "loss": 0.1082,
+      "step": 2620
+    },
+    {
+      "epoch": 2.846320346320346,
+      "grad_norm": 0.28515625,
+      "learning_rate": 2.6571856287425152e-05,
+      "loss": 0.1125,
+      "step": 2630
+    },
+    {
+      "epoch": 2.857142857142857,
+      "grad_norm": 0.07568359375,
+      "learning_rate": 2.4700598802395212e-05,
+      "loss": 0.0981,
+      "step": 2640
+    },
+    {
+      "epoch": 2.867965367965368,
+      "grad_norm": 0.169921875,
+      "learning_rate": 2.282934131736527e-05,
+      "loss": 0.1197,
+      "step": 2650
+    },
+    {
+      "epoch": 2.878787878787879,
+      "grad_norm": 0.34765625,
+      "learning_rate": 2.095808383233533e-05,
+      "loss": 0.1214,
+      "step": 2660
+    },
+    {
+      "epoch": 2.8896103896103895,
+      "grad_norm": 0.1513671875,
+      "learning_rate": 1.9086826347305392e-05,
+      "loss": 0.105,
+      "step": 2670
+    },
+    {
+      "epoch": 2.9004329004329006,
+      "grad_norm": 0.1767578125,
+      "learning_rate": 1.7215568862275448e-05,
+      "loss": 0.1101,
+      "step": 2680
+    },
+    {
+      "epoch": 2.911255411255411,
+      "grad_norm": 0.169921875,
+      "learning_rate": 1.5344311377245508e-05,
+      "loss": 0.1283,
+      "step": 2690
+    },
+    {
+      "epoch": 2.9220779220779223,
+      "grad_norm": 0.130859375,
+      "learning_rate": 1.347305389221557e-05,
+      "loss": 0.1126,
+      "step": 2700
+    },
+    {
+      "epoch": 2.9220779220779223,
+      "eval_loss": 0.1572347730398178,
+      "eval_runtime": 58.0947,
+      "eval_samples_per_second": 8.607,
+      "eval_steps_per_second": 0.069,
+      "step": 2700
+    },
+    {
+      "epoch": 2.932900432900433,
+      "grad_norm": 0.197265625,
+      "learning_rate": 1.1601796407185628e-05,
+      "loss": 0.1127,
+      "step": 2710
+    },
+    {
+      "epoch": 2.9437229437229435,
+      "grad_norm": 0.259765625,
+      "learning_rate": 9.73053892215569e-06,
+      "loss": 0.1236,
+      "step": 2720
+    },
+    {
+      "epoch": 2.9545454545454546,
+      "grad_norm": 0.1416015625,
+      "learning_rate": 7.859281437125748e-06,
+      "loss": 0.1137,
+      "step": 2730
+    },
+    {
+      "epoch": 2.965367965367965,
+      "grad_norm": 0.2041015625,
+      "learning_rate": 5.9880239520958085e-06,
+      "loss": 0.1213,
+      "step": 2740
+    },
+    {
+      "epoch": 2.9761904761904763,
+      "grad_norm": 0.1728515625,
+      "learning_rate": 4.116766467065868e-06,
+      "loss": 0.1083,
+      "step": 2750
+    },
+    {
+      "epoch": 2.987012987012987,
+      "grad_norm": 0.2099609375,
+      "learning_rate": 2.2455089820359283e-06,
+      "loss": 0.1217,
+      "step": 2760
+    },
+    {
+      "epoch": 2.997835497835498,
+      "grad_norm": 0.2099609375,
+      "learning_rate": 3.7425149700598803e-07,
+      "loss": 0.1188,
+      "step": 2770
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 2772,
+  "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": 6.887405812854188e+18,
+  "train_batch_size": 128,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-2772/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f2afcf823dba854c82182f2115a254b2688cf06dceefd5def0595b5a4da45c35
+size 5240

git_hash.txt

@@ -0,0 +1 @@
+4585c16aaf819f24684c118bdacfbf1514bb56b9

preprocessor_config.json

@@ -0,0 +1,28 @@
+{
+  "do_convert_rgb": true,
+  "do_image_splitting": true,
+  "do_normalize": true,
+  "do_pad": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "Idefics3ImageProcessor",
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "max_image_size": {
+    "longest_edge": 384
+  },
+  "processor_class": "ColIdefics3Processor",
+  "resample": 1,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "longest_edge": 1536
+  }
+}

results.json

@@ -0,0 +1 @@
+{"validation_set": {"ndcg_at_1": 0.828, "ndcg_at_3": 0.87764, "ndcg_at_5": 0.88263, "ndcg_at_10": 0.88981, "ndcg_at_20": 0.89434, "ndcg_at_50": 0.89829, "ndcg_at_100": 0.90018, "map_at_1": 0.828, "map_at_3": 0.86633, "map_at_5": 0.86913, "map_at_10": 0.87213, "map_at_20": 0.87336, "map_at_50": 0.87399, "map_at_100": 0.87413, "recall_at_1": 0.828, "recall_at_3": 0.91, "recall_at_5": 0.922, "recall_at_10": 0.944, "recall_at_20": 0.962, "recall_at_50": 0.982, "recall_at_100": 0.994, "precision_at_1": 0.828, "precision_at_3": 0.30333, "precision_at_5": 0.1844, "precision_at_10": 0.0944, "precision_at_20": 0.0481, "precision_at_50": 0.01964, "precision_at_100": 0.00994, "mrr_at_1": 0.824, "mrr_at_3": 0.8636666666666666, "mrr_at_5": 0.8669666666666667, "mrr_at_10": 0.8702642857142857, "mrr_at_20": 0.871532798573975, "mrr_at_50": 0.8722038761027637, "mrr_at_100": 0.872391544636181, "naucs_at_1_max": 0.21773711896722028, "naucs_at_1_std": 0.2788401902005385, "naucs_at_1_diff1": 0.9399880550387059, "naucs_at_3_max": 0.15050316422865537, "naucs_at_3_std": 0.6578586990351711, "naucs_at_3_diff1": 0.9553376906318075, "naucs_at_5_max": 0.1060954296248405, "naucs_at_5_std": 0.7918073212190876, "naucs_at_5_diff1": 0.9656443774090842, "naucs_at_10_max": -0.018807523009205383, "naucs_at_10_std": 0.86072762438309, "naucs_at_10_diff1": 0.9807422969187695, "naucs_at_20_max": -0.3403607056857979, "naucs_at_20_std": 0.9314708339476163, "naucs_at_20_diff1": 1.0, "naucs_at_50_max": -0.5622471210706596, "naucs_at_50_std": 1.0, "naucs_at_50_diff1": 1.0, "naucs_at_100_max": -0.6288515406162857, "naucs_at_100_std": 1.0, "naucs_at_100_diff1": 1.0}, "syntheticDocQA_energy": {"ndcg_at_1": 0.94, "ndcg_at_3": 0.97155, "ndcg_at_5": 0.97155, "ndcg_at_10": 0.97155, "ndcg_at_20": 0.97155, "ndcg_at_50": 0.9735, "ndcg_at_100": 0.9735, "map_at_1": 0.94, "map_at_3": 0.965, "map_at_5": 0.965, "map_at_10": 0.965, "map_at_20": 0.965, "map_at_50": 0.96529, "map_at_100": 0.96529, "recall_at_1": 0.94, "recall_at_3": 0.99, "recall_at_5": 0.99, "recall_at_10": 0.99, "recall_at_20": 0.99, "recall_at_50": 1.0, "recall_at_100": 1.0, "precision_at_1": 0.94, "precision_at_3": 0.33, "precision_at_5": 0.198, "precision_at_10": 0.099, "precision_at_20": 0.0495, "precision_at_50": 0.02, "precision_at_100": 0.01, "mrr_at_1": 0.95, "mrr_at_3": 0.97, "mrr_at_5": 0.97, "mrr_at_10": 0.97, "mrr_at_20": 0.97, "mrr_at_50": 0.9702941176470589, "mrr_at_100": 0.9702941176470589, "naucs_at_1_max": 0.5834111422346712, "naucs_at_1_std": -0.3165266106442549, "naucs_at_1_diff1": 0.9782135076252712, "naucs_at_3_max": 0.7222222222222157, "naucs_at_3_std": -0.5634920634921204, "naucs_at_3_diff1": 1.0, "naucs_at_5_max": 0.7222222222222276, "naucs_at_5_std": -0.5634920634920767, "naucs_at_5_diff1": 1.0, "naucs_at_10_max": 0.7222222222222276, "naucs_at_10_std": -0.5634920634920767, "naucs_at_10_diff1": 1.0, "naucs_at_20_max": 0.7222222222222276, "naucs_at_20_std": -0.5634920634920767, "naucs_at_20_diff1": 1.0, "naucs_at_50_max": NaN, "naucs_at_50_std": NaN, "naucs_at_50_diff1": NaN, "naucs_at_100_max": NaN, "naucs_at_100_std": NaN, "naucs_at_100_diff1": NaN}, "syntheticDocQA_healthcare_industry": {"ndcg_at_1": 0.98, "ndcg_at_3": 0.99262, "ndcg_at_5": 0.99262, "ndcg_at_10": 0.99262, "ndcg_at_20": 0.99262, "ndcg_at_50": 0.99262, "ndcg_at_100": 0.99262, "map_at_1": 0.98, "map_at_3": 0.99, "map_at_5": 0.99, "map_at_10": 0.99, "map_at_20": 0.99, "map_at_50": 0.99, "map_at_100": 0.99, "recall_at_1": 0.98, "recall_at_3": 1.0, "recall_at_5": 1.0, "recall_at_10": 1.0, "recall_at_20": 1.0, "recall_at_50": 1.0, "recall_at_100": 1.0, "precision_at_1": 0.98, "precision_at_3": 0.33333, "precision_at_5": 0.2, "precision_at_10": 0.1, "precision_at_20": 0.05, "precision_at_50": 0.02, "precision_at_100": 0.01, "mrr_at_1": 0.98, "mrr_at_3": 0.99, "mrr_at_5": 0.99, "mrr_at_10": 0.99, "mrr_at_20": 0.99, "mrr_at_50": 0.99, "mrr_at_100": 0.99, "naucs_at_1_max": 0.795751633986929, "naucs_at_1_std": -0.8576097105508799, "naucs_at_1_diff1": 0.9346405228758133, "naucs_at_3_max": 1.0, "naucs_at_3_std": 1.0, "naucs_at_3_diff1": 1.0, "naucs_at_5_max": 1.0, "naucs_at_5_std": 1.0, "naucs_at_5_diff1": 1.0, "naucs_at_10_max": 1.0, "naucs_at_10_std": 1.0, "naucs_at_10_diff1": 1.0, "naucs_at_20_max": 1.0, "naucs_at_20_std": 1.0, "naucs_at_20_diff1": 1.0, "naucs_at_50_max": NaN, "naucs_at_50_std": NaN, "naucs_at_50_diff1": NaN, "naucs_at_100_max": NaN, "naucs_at_100_std": NaN, "naucs_at_100_diff1": NaN}, "syntheticDocQA_artificial_intelligence_test": {"ndcg_at_1": 0.99, "ndcg_at_3": 0.995, "ndcg_at_5": 0.995, "ndcg_at_10": 0.995, "ndcg_at_20": 0.995, "ndcg_at_50": 0.995, "ndcg_at_100": 0.995, "map_at_1": 0.99, "map_at_3": 0.99333, "map_at_5": 0.99333, "map_at_10": 0.99333, "map_at_20": 0.99333, "map_at_50": 0.99333, "map_at_100": 0.99333, "recall_at_1": 0.99, "recall_at_3": 1.0, "recall_at_5": 1.0, "recall_at_10": 1.0, "recall_at_20": 1.0, "recall_at_50": 1.0, "recall_at_100": 1.0, "precision_at_1": 0.99, "precision_at_3": 0.33333, "precision_at_5": 0.2, "precision_at_10": 0.1, "precision_at_20": 0.05, "precision_at_50": 0.02, "precision_at_100": 0.01, "mrr_at_1": 0.99, "mrr_at_3": 0.9933333333333333, "mrr_at_5": 0.9933333333333333, "mrr_at_10": 0.9933333333333333, "mrr_at_20": 0.9933333333333333, "mrr_at_50": 0.9933333333333333, "mrr_at_100": 0.9933333333333333, "naucs_at_1_max": 1.0, "naucs_at_1_std": 0.8692810457516276, "naucs_at_1_diff1": 1.0, "naucs_at_3_max": 1.0, "naucs_at_3_std": 1.0, "naucs_at_3_diff1": 1.0, "naucs_at_5_max": 1.0, "naucs_at_5_std": 1.0, "naucs_at_5_diff1": 1.0, "naucs_at_10_max": 1.0, "naucs_at_10_std": 1.0, "naucs_at_10_diff1": 1.0, "naucs_at_20_max": 1.0, "naucs_at_20_std": 1.0, "naucs_at_20_diff1": 1.0, "naucs_at_50_max": NaN, "naucs_at_50_std": NaN, "naucs_at_50_diff1": NaN, "naucs_at_100_max": NaN, "naucs_at_100_std": NaN, "naucs_at_100_diff1": NaN}, "syntheticDocQA_government_reports": {"ndcg_at_1": 0.93, "ndcg_at_3": 0.97155, "ndcg_at_5": 0.97155, "ndcg_at_10": 0.97155, "ndcg_at_20": 0.97155, "ndcg_at_50": 0.97155, "ndcg_at_100": 0.97155, "map_at_1": 0.93, "map_at_3": 0.96167, "map_at_5": 0.96167, "map_at_10": 0.96167, "map_at_20": 0.96167, "map_at_50": 0.96167, "map_at_100": 0.96167, "recall_at_1": 0.93, "recall_at_3": 1.0, "recall_at_5": 1.0, "recall_at_10": 1.0, "recall_at_20": 1.0, "recall_at_50": 1.0, "recall_at_100": 1.0, "precision_at_1": 0.93, "precision_at_3": 0.33333, "precision_at_5": 0.2, "precision_at_10": 0.1, "precision_at_20": 0.05, "precision_at_50": 0.02, "precision_at_100": 0.01, "mrr_at_1": 0.94, "mrr_at_3": 0.9666666666666667, "mrr_at_5": 0.9666666666666667, "mrr_at_10": 0.9666666666666667, "mrr_at_20": 0.9666666666666667, "mrr_at_50": 0.9666666666666667, "mrr_at_100": 0.9666666666666667, "naucs_at_1_max": 0.5822328931572631, "naucs_at_1_std": 0.673536081099108, "naucs_at_1_diff1": 0.8622782446311864, "naucs_at_3_max": 1.0, "naucs_at_3_std": 1.0, "naucs_at_3_diff1": 1.0, "naucs_at_5_max": 1.0, "naucs_at_5_std": 1.0, "naucs_at_5_diff1": 1.0, "naucs_at_10_max": 1.0, "naucs_at_10_std": 1.0, "naucs_at_10_diff1": 1.0, "naucs_at_20_max": 1.0, "naucs_at_20_std": 1.0, "naucs_at_20_diff1": 1.0, "naucs_at_50_max": NaN, "naucs_at_50_std": NaN, "naucs_at_50_diff1": NaN, "naucs_at_100_max": NaN, "naucs_at_100_std": NaN, "naucs_at_100_diff1": NaN}, "infovqa_subsampled": {"ndcg_at_1": 0.87, "ndcg_at_3": 0.89567, "ndcg_at_5": 0.9047, "ndcg_at_10": 0.90993, "ndcg_at_20": 0.91299, "ndcg_at_50": 0.91631, "ndcg_at_100": 0.91664, "map_at_1": 0.87, "map_at_3": 0.88933, "map_at_5": 0.89433, "map_at_10": 0.89652, "map_at_20": 0.89737, "map_at_50": 0.89796, "map_at_100": 0.89799, "recall_at_1": 0.87, "recall_at_3": 0.914, "recall_at_5": 0.936, "recall_at_10": 0.952, "recall_at_20": 0.964, "recall_at_50": 0.98, "recall_at_100": 0.982, "precision_at_1": 0.87, "precision_at_3": 0.30467, "precision_at_5": 0.1872, "precision_at_10": 0.0952, "precision_at_20": 0.0482, "precision_at_50": 0.0196, "precision_at_100": 0.00982, "mrr_at_1": 0.872, "mrr_at_3": 0.8903333333333334, "mrr_at_5": 0.8952333333333333, "mrr_at_10": 0.8973857142857142, "mrr_at_20": 0.8981114902641218, "mrr_at_50": 0.8987658806482948, "mrr_at_100": 0.8987961836785979, "naucs_at_1_max": 0.6437088054506402, "naucs_at_1_std": 0.05869806709620076, "naucs_at_1_diff1": 0.9304969266089019, "naucs_at_3_max": 0.6889344885240949, "naucs_at_3_std": 0.1136842333832744, "naucs_at_3_diff1": 0.9251405988752106, "naucs_at_5_max": 0.7031395891690008, "naucs_at_5_std": 0.26028536414565784, "naucs_at_5_diff1": 0.9203431372549006, "naucs_at_10_max": 0.8295012449424186, "naucs_at_10_std": 0.40411998132586524, "naucs_at_10_diff1": 0.9332788671023946, "naucs_at_20_max": 0.8841944185081392, "naucs_at_20_std": 0.49084448594252705, "naucs_at_20_diff1": 0.9337327523601999, "naucs_at_50_max": 0.9292717086834656, "naucs_at_50_std": 0.8650793650793516, "naucs_at_50_diff1": 0.9477124183006472, "naucs_at_100_max": 0.9214130096482952, "naucs_at_100_std": 0.8646125116713362, "naucs_at_100_diff1": 0.9419026870007192}, "docvqa_subsampled": {"ndcg_at_1": 0.478, "ndcg_at_3": 0.53969, "ndcg_at_5": 0.56094, "ndcg_at_10": 0.57779, "ndcg_at_20": 0.59308, "ndcg_at_50": 0.60319, "ndcg_at_100": 0.60994, "map_at_1": 0.478, "map_at_3": 0.525, "map_at_5": 0.5367, "map_at_10": 0.54367, "map_at_20": 0.54794, "map_at_50": 0.54964, "map_at_100": 0.55022, "recall_at_1": 0.478, "recall_at_3": 0.582, "recall_at_5": 0.634, "recall_at_10": 0.686, "recall_at_20": 0.746, "recall_at_50": 0.796, "recall_at_100": 0.838, "precision_at_1": 0.478, "precision_at_3": 0.194, "precision_at_5": 0.1268, "precision_at_10": 0.0686, "precision_at_20": 0.0373, "precision_at_50": 0.01592, "precision_at_100": 0.00838, "mrr_at_1": 0.482, "mrr_at_3": 0.5266666666666666, "mrr_at_5": 0.5375666666666667, "mrr_at_10": 0.5453944444444444, "mrr_at_20": 0.5492540835755232, "mrr_at_50": 0.5511465806213975, "mrr_at_100": 0.5515877171220239, "naucs_at_1_max": 0.17593079866611316, "naucs_at_1_std": 0.6646824085068347, "naucs_at_1_diff1": 0.8815686129321448, "naucs_at_3_max": 0.08521158919497512, "naucs_at_3_std": 0.7560230713519073, "naucs_at_3_diff1": 0.7969302187589984, "naucs_at_5_max": 0.02873191231399215, "naucs_at_5_std": 0.8071899242551022, "naucs_at_5_diff1": 0.7628796322315227, "naucs_at_10_max": -0.07012897967179643, "naucs_at_10_std": 0.874808909287446, "naucs_at_10_diff1": 0.7255395142360395, "naucs_at_20_max": -0.20666283697498458, "naucs_at_20_std": 0.9340741995875509, "naucs_at_20_diff1": 0.6769916807524927, "naucs_at_50_max": -0.25364250789497006, "naucs_at_50_std": 0.913686581994549, "naucs_at_50_diff1": 0.6194624533926231, "naucs_at_100_max": -0.33117544987926306, "naucs_at_100_std": 0.925380107024548, "naucs_at_100_diff1": 0.5889383698778079}, "arxivqa_subsampled": {"ndcg_at_1": 0.774, "ndcg_at_3": 0.82333, "ndcg_at_5": 0.83168, "ndcg_at_10": 0.84763, "ndcg_at_20": 0.85417, "ndcg_at_50": 0.86064, "ndcg_at_100": 0.86196, "map_at_1": 0.774, "map_at_3": 0.81133, "map_at_5": 0.81603, "map_at_10": 0.82249, "map_at_20": 0.82428, "map_at_50": 0.82536, "map_at_100": 0.82548, "recall_at_1": 0.774, "recall_at_3": 0.858, "recall_at_5": 0.878, "recall_at_10": 0.928, "recall_at_20": 0.954, "recall_at_50": 0.986, "recall_at_100": 0.994, "precision_at_1": 0.774, "precision_at_3": 0.286, "precision_at_5": 0.1756, "precision_at_10": 0.0928, "precision_at_20": 0.0477, "precision_at_50": 0.01972, "precision_at_100": 0.00994, "mrr_at_1": 0.774, "mrr_at_3": 0.8103333333333333, "mrr_at_5": 0.8154333333333332, "mrr_at_10": 0.8216412698412698, "mrr_at_20": 0.8235997861185941, "mrr_at_50": 0.8245970376853287, "mrr_at_100": 0.824717741545969, "naucs_at_1_max": 0.6956030338929966, "naucs_at_1_std": 0.034737257785584987, "naucs_at_1_diff1": 0.9216607432220811, "naucs_at_3_max": 0.7191420090012451, "naucs_at_3_std": 0.10592878346397475, "naucs_at_3_diff1": 0.8826691837320956, "naucs_at_5_max": 0.7499255380853105, "naucs_at_5_std": 0.12009531125080068, "naucs_at_5_diff1": 0.8803514602373368, "naucs_at_10_max": 0.8015224608361871, "naucs_at_10_std": 0.08101203444340487, "naucs_at_10_diff1": 0.8493749351592488, "naucs_at_20_max": 0.7641375390735974, "naucs_at_20_std": 0.1359355336337385, "naucs_at_20_diff1": 0.8352819388625017, "naucs_at_50_max": 0.7675736961451084, "naucs_at_50_std": 0.057823129251703305, "naucs_at_50_diff1": 0.8646125116713221, "naucs_at_100_max": 0.7424525365701481, "naucs_at_100_std": 0.49361967009023777, "naucs_at_100_diff1": 0.8638344226579416}, "tabfquad_subsampled": {"ndcg_at_1": 0.73571, "ndcg_at_3": 0.79375, "ndcg_at_5": 0.80251, "ndcg_at_10": 0.81636, "ndcg_at_20": 0.82817, "ndcg_at_50": 0.83832, "ndcg_at_100": 0.83954, "map_at_1": 0.73571, "map_at_3": 0.78036, "map_at_5": 0.78518, "map_at_10": 0.7909, "map_at_20": 0.79416, "map_at_50": 0.79589, "map_at_100": 0.79601, "recall_at_1": 0.73571, "recall_at_3": 0.83214, "recall_at_5": 0.85357, "recall_at_10": 0.89643, "recall_at_20": 0.94286, "recall_at_50": 0.99286, "recall_at_100": 1.0, "precision_at_1": 0.73571, "precision_at_3": 0.27738, "precision_at_5": 0.17071, "precision_at_10": 0.08964, "precision_at_20": 0.04714, "precision_at_50": 0.01986, "precision_at_100": 0.01, "mrr_at_1": 0.7357142857142858, "mrr_at_3": 0.7797619047619048, "mrr_at_5": 0.7838690476190475, "mrr_at_10": 0.7901204648526078, "mrr_at_20": 0.7934665929308786, "mrr_at_50": 0.7952056267189848, "mrr_at_100": 0.7953287794283443, "naucs_at_1_max": 0.3253261692793548, "naucs_at_1_std": 0.06312608068134669, "naucs_at_1_diff1": 0.8084496777604704, "naucs_at_3_max": 0.3105270885332547, "naucs_at_3_std": 0.07538225978983691, "naucs_at_3_diff1": 0.738448308408667, "naucs_at_5_max": 0.31633840963626375, "naucs_at_5_std": 0.09126654154094667, "naucs_at_5_diff1": 0.7328288707799767, "naucs_at_10_max": 0.1824121090591562, "naucs_at_10_std": -0.010980303373330084, "naucs_at_10_diff1": 0.6897377017367963, "naucs_at_20_max": 0.0169234360410811, "naucs_at_20_std": -0.31098272642390185, "naucs_at_20_diff1": 0.6254668534080327, "naucs_at_50_max": -0.3968253968254018, "naucs_at_50_std": -0.690943043884227, "naucs_at_50_diff1": 0.8611111111110865, "naucs_at_100_max": 1.0, "naucs_at_100_std": 1.0, "naucs_at_100_diff1": 1.0}, "tatdqa": {"ndcg_at_1": 0.66987, "ndcg_at_3": 0.77553, "ndcg_at_5": 0.79538, "ndcg_at_10": 0.80853, "ndcg_at_20": 0.81354, "ndcg_at_50": 0.81796, "ndcg_at_100": 0.81903, "map_at_1": 0.66987, "map_at_3": 0.75005, "map_at_5": 0.76108, "map_at_10": 0.76669, "map_at_20": 0.76814, "map_at_50": 0.76886, "map_at_100": 0.76895, "recall_at_1": 0.66987, "recall_at_3": 0.84907, "recall_at_5": 0.89717, "recall_at_10": 0.93686, "recall_at_20": 0.9561, "recall_at_50": 0.97835, "recall_at_100": 0.98497, "precision_at_1": 0.66987, "precision_at_3": 0.28302, "precision_at_5": 0.17943, "precision_at_10": 0.09369, "precision_at_20": 0.04781, "precision_at_50": 0.01957, "precision_at_100": 0.00985, "mrr_at_1": 0.6650631389055923, "mrr_at_3": 0.7486470234515935, "mrr_at_5": 0.7588995790739627, "mrr_at_10": 0.7649008771678646, "mrr_at_20": 0.766294670223957, "mrr_at_50": 0.7669629679577795, "mrr_at_100": 0.7670583172691432, "naucs_at_1_max": 0.2568730737793943, "naucs_at_1_std": -0.12808142195797126, "naucs_at_1_diff1": 0.8234815744151925, "naucs_at_3_max": 0.32424305760807287, "naucs_at_3_std": 0.0023177179676344157, "naucs_at_3_diff1": 0.7226885160746074, "naucs_at_5_max": 0.3697612224125756, "naucs_at_5_std": 0.11798240403444073, "naucs_at_5_diff1": 0.6833142525048825, "naucs_at_10_max": 0.49587555962938135, "naucs_at_10_std": 0.22528738928970488, "naucs_at_10_diff1": 0.7079879601645019, "naucs_at_20_max": 0.5475720475651525, "naucs_at_20_std": 0.3970102434994993, "naucs_at_20_diff1": 0.7067977425806262, "naucs_at_50_max": 0.6739875444992179, "naucs_at_50_std": 0.5159845368869518, "naucs_at_50_diff1": 0.7293116370600532, "naucs_at_100_max": 0.7557825589788986, "naucs_at_100_std": 0.5991543861889267, "naucs_at_100_diff1": 0.787492165390262}, "shift_project": {"ndcg_at_1": 0.67, "ndcg_at_3": 0.77464, "ndcg_at_5": 0.80778, "ndcg_at_10": 0.81467, "ndcg_at_20": 0.82248, "ndcg_at_50": 0.82621, "ndcg_at_100": 0.82621, "map_at_1": 0.67, "map_at_3": 0.75167, "map_at_5": 0.77017, "map_at_10": 0.77326, "map_at_20": 0.77553, "map_at_50": 0.77603, "map_at_100": 0.77603, "recall_at_1": 0.67, "recall_at_3": 0.84, "recall_at_5": 0.92, "recall_at_10": 0.94, "recall_at_20": 0.97, "recall_at_50": 0.99, "recall_at_100": 0.99, "precision_at_1": 0.67, "precision_at_3": 0.28, "precision_at_5": 0.184, "precision_at_10": 0.094, "precision_at_20": 0.0485, "precision_at_50": 0.0198, "precision_at_100": 0.0099, "mrr_at_1": 0.71, "mrr_at_3": 0.775, "mrr_at_5": 0.7935, "mrr_at_10": 0.7961666666666666, "mrr_at_20": 0.7976639928698752, "mrr_at_50": 0.7981844144772533, "mrr_at_100": 0.7981844144772533, "naucs_at_1_max": 0.24121819505341022, "naucs_at_1_std": -0.1675272130148469, "naucs_at_1_diff1": 0.8373652045843135, "naucs_at_3_max": 0.26165930289641504, "naucs_at_3_std": -0.26902307314678475, "naucs_at_3_diff1": 0.697379725085911, "naucs_at_5_max": -0.003501400560219899, "naucs_at_5_std": -0.6327614379084904, "naucs_at_5_diff1": 0.5134220354808603, "naucs_at_10_max": 0.02956738250855913, "naucs_at_10_std": -0.623171490818547, "naucs_at_10_diff1": 0.5325241207594161, "naucs_at_20_max": -0.20572673513850018, "naucs_at_20_std": -1.1190476190476195, "naucs_at_20_diff1": 0.6150015561780299, "naucs_at_50_max": 0.12278244631185525, "naucs_at_50_std": 0.12278244631185525, "naucs_at_50_diff1": 0.12278244631185525, "naucs_at_100_max": 0.12278244631185525, "naucs_at_100_std": 0.12278244631185525, "naucs_at_100_diff1": 0.12278244631185525}}

special_tokens_map.json

@@ -0,0 +1,53 @@
+{
+  "additional_special_tokens": [
+    {
+      "content": "<fake_token_around_image>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<image>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<end_of_utterance>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
+  "bos_token": {
+    "content": "<|im_start|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<end_of_utterance>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer_config.json

@@ -0,0 +1,182 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "<repo_name>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "4": {
+      "content": "<reponame>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "5": {
+      "content": "<file_sep>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "6": {
+      "content": "<filename>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "7": {
+      "content": "<gh_stars>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "8": {
+      "content": "<issue_start>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "9": {
+      "content": "<issue_comment>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "10": {
+      "content": "<issue_closed>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "11": {
+      "content": "<jupyter_start>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "12": {
+      "content": "<jupyter_text>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "13": {
+      "content": "<jupyter_code>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "14": {
+      "content": "<jupyter_output>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "15": {
+      "content": "<jupyter_script>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "16": {
+      "content": "<empty_output>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "49152": {
+      "content": "<fake_token_around_image>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "49153": {
+      "content": "<image>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "49154": {
+      "content": "<end_of_utterance>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<fake_token_around_image>",
+    "<image>",
+    "<end_of_utterance>"
+  ],
+  "bos_token": "<|im_start|>",
+  "chat_template": "{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<end_of_utterance>",
+  "legacy": false,
+  "model_max_length": 16384,
+  "pad_token": "<|im_end|>",
+  "processor_class": "ColIdefics3Processor",
+  "tokenizer_class": "GPT2Tokenizer",
+  "truncation_side": "left",
+  "unk_token": "<|endoftext|>",
+  "vocab_size": 49152
+}

training_config.yml

@@ -0,0 +1,72 @@
+config:
+  (): colpali_engine.trainer.colmodel_training.ColModelTrainingConfig
+  output_dir: !path ../../../models/colsmolvlm-b64
+  processor:
+    (): colpali_engine.utils.transformers_wrappers.AllPurposeWrapper
+    class_to_instanciate: !ext colpali_engine.models.ColIdefics3Processor
+    pretrained_model_name_or_path:  "./models/SmolVLM-Instruct" # "./models/paligemma-3b-mix-448"
+    # num_image_tokens: 2048
+    # max_length: 50
+
+  model:
+    (): colpali_engine.utils.transformers_wrappers.AllPurposeWrapper
+    class_to_instanciate: !ext colpali_engine.models.ColIdefics3
+    pretrained_model_name_or_path: "./models/SmolVLM-Instruct"
+    torch_dtype:  !ext torch.bfloat16
+    # use_cache: false
+    attn_implementation: "flash_attention_2"
+#    device_map: "auto"
+#    quantization_config:
+#      (): transformers.BitsAndBytesConfig
+#      load_in_4bit: true
+#      bnb_4bit_quant_type: "nf4"
+#      bnb_4bit_compute_dtype:  "bfloat16"
+#      bnb_4bit_use_double_quant: true
+
+  dataset_loading_func: !ext colpali_engine.utils.dataset_transformation.load_train_set
+  eval_dataset_loader: !import ../data/test_data.yaml
+
+  # max_length: 50
+  run_eval: true
+  loss_func:
+    (): colpali_engine.loss.late_interaction_losses.ColbertPairwiseCELoss
+  tr_args:
+    (): transformers.training_args.TrainingArguments
+    output_dir: null
+    overwrite_output_dir: true
+    num_train_epochs: 3
+    per_device_train_batch_size: 32
+    gradient_checkpointing: true
+    gradient_checkpointing_kwargs: { "use_reentrant": false }
+    # gradient_checkpointing: true
+    # 6 x 8 gpus = 48 batch size
+    # gradient_accumulation_steps: 4
+    per_device_eval_batch_size: 32
+    eval_strategy: "steps"
+    dataloader_num_workers: 8
+    # bf16: true
+    save_steps: 500
+    logging_steps: 10
+    eval_steps: 100
+    warmup_steps: 100
+    learning_rate: 5e-4
+    save_total_limit: 1
+    # resume_from_checkpoint: true
+    # optim: "paged_adamw_8bit"
+    # wandb logging
+    # wandb_project: "colqwen2"
+    # run_name: "colqwen2-ba32-nolora"
+    report_to: "wandb"
+
+
+  peft_config:
+    (): peft.LoraConfig
+    r: 32
+    lora_alpha: 32
+    lora_dropout: 0.1
+    init_lora_weights: "gaussian"
+    bias: "none"
+    task_type: "FEATURE_EXTRACTION"
+    target_modules: '(.*(model.text_model).*(down_proj|gate_proj|up_proj|k_proj|q_proj|v_proj|o_proj).*$|.*(custom_text_proj).*$)'
+    # target_modules: '(.*(language_model).*(down_proj|gate_proj|up_proj|k_proj|q_proj|v_proj|o_proj).*$|.*(custom_text_proj).*$)'
+