Add files using upload-large-folder tool

.gitattributes

@@ -33,3 +33,9 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+examples/pyimage.png filter=lfs diff=lfs merge=lfs -text
+examples/tikzimage.png filter=lfs diff=lfs merge=lfs -text
+examples/tab1.png filter=lfs diff=lfs merge=lfs -text
+examples/fig1.png filter=lfs diff=lfs merge=lfs -text
+examples/fig2.png filter=lfs diff=lfs merge=lfs -text
+examples/fig4.png filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,3 +1,110 @@
----
-license: apache-2.0
----
+---
+license: apache-2.0
+language:
+- en
+metrics:
+- accuracy
+pipeline_tag: text-generation
+---
+# MathCoder-VL: Bridging Vision and Code for Enhanced Multimodal Mathematical Reasoning
+
+Repo: [https://github.com/mathllm/MathCoder](https://github.com/mathllm/MathCoder)
+
+
+## Introduction
+We introduce MathCoder-VL, a series of open-source large multimodal models (LMMs) specifically tailored for general math problem-solving. We also introduce [FigCodifier-8B](https://huggingface.co/MathLLMs/FigCodifier), an image-to-code model.
+
+| Base Model                                          	|Ours                                               |
+|-------------------------------------------------------------------|-----------------------------------------------------------------------|
+|  [Mini-InternVL-Chat-2B-V1-5](https://huggingface.co/OpenGVLab/Mini-InternVL-Chat-2B-V1-5)  |  [MathCoder-VL-2B](https://huggingface.co/MathLLMs/MathCoder-VL-2B)   	|
+|  [InternVL2-8B](https://huggingface.co/OpenGVLab/InternVL2-8B)  |     	[MathCoder-VL-8B](https://huggingface.co/MathLLMs/MathCoder-VL-8B)|
+
+
+
+## Usage
+For training and inference code, please refer to [InternVL](https://github.com/OpenGVLab/InternVL).
+
+### Prompt for TikZ Code Generation
+
+```
+<image>\nPlease generate the corresponding TikZ code that accurately represents the visual elements in the image. TikZ is a powerful tool for creating vector graphics within LaTeX documents. Your generated code should be precise, well-structured, and should recreate the image as faithfully as possible.
+```
+
+<div align="center">
+  <img src="./examples/tikzimage.png" width="100%" title="Result Figure">
+</div>
+
+### Prompt for Python Code Generation
+
+```
+Please provide the Python code needed to reproduce this image.\n<image>
+<div align="center">
+  <img src="./examples/pyimage.png" width="100%" title="Result Figure">
+</div>
+```
+
+<div align="center">
+  <img src="./examples/pyimage.png" width="100%" title="Result Figure">
+</div>
+
+
+## Motivation
+<br>
+<div align="center">
+  <img src="./examples/fig1.png" width="100%" title="Result Figure">
+</div>
+
+## Construction of FigCodifier
+
+<br>
+<div align="center">
+  <img src="./examples/fig2.png" width="100%" title="Result Figure">
+</div>
+
+## Construction of MathCoder-VL
+
+
+<div align="center">
+  <img src="./examples/fig4.png" width="100%" title="Result Figure">
+</div>
+
+## Performance
+
+<div align="center">
+  <img src="./examples/tab1.png" width="100%" title="Result Figure">
+</div>
+
+## **Citation**
+
+Please cite the paper if you use our data, model or code.
+
+```
+@inproceedings{
+wang2025mathcodervl,
+title={MathCoder-{VL}: Bridging Vision and Code for Enhanced Multimodal Mathematical Reasoning},
+author={Ke Wang and Junting Pan and Linda Wei and Aojun Zhou and Weikang Shi and Zimu Lu and Han Xiao and Yunqiao Yang and Houxing Ren and Mingjie Zhan and Hongsheng Li},
+booktitle={The 63rd Annual Meeting of the Association for Computational Linguistics},
+year={2025},
+url={https://openreview.net/forum?id=nuvtX1imAb}
+}
+```
+```
+@inproceedings{
+lu2025mathcoder,
+title={MathCoder2: Better Math Reasoning from Continued Pretraining on Model-translated Mathematical Code},
+author={Zimu Lu and Aojun Zhou and Ke Wang and Houxing Ren and Weikang Shi and Junting Pan and Mingjie Zhan and Hongsheng Li},
+booktitle={The Thirteenth International Conference on Learning Representations},
+year={2025},
+url={https://openreview.net/forum?id=1Iuw1jcIrf}
+}
+```
+```
+@inproceedings{
+wang2024mathcoder,
+title={MathCoder: Seamless Code Integration in {LLM}s for Enhanced Mathematical Reasoning},
+author={Ke Wang and Houxing Ren and Aojun Zhou and Zimu Lu and Sichun Luo and Weikang Shi and Renrui Zhang and Linqi Song and Mingjie Zhan and Hongsheng Li},
+booktitle={The Twelfth International Conference on Learning Representations},
+year={2024},
+url={https://openreview.net/forum?id=z8TW0ttBPp}
+}
+```

added_tokens.json

@@ -0,0 +1,15 @@
+{
+  "</box>": 92552,
+  "</img>": 92545,
+  "</quad>": 92548,
+  "</ref>": 92550,
+  "<IMG_CONTEXT>": 92546,
+  "<box>": 92551,
+  "<img>": 92544,
+  "<quad>": 92547,
+  "<ref>": 92549,
+  "<|/execution|>": 92556,
+  "<|/python|>": 92554,
+  "<|execution|>": 92555,
+  "<|python|>": 92553
+}

config.json

@@ -0,0 +1,201 @@
+{
+  "_commit_hash": null,
+  "_name_or_path": "./pretrained/InternVL2-8B",
+  "architectures": [
+    "InternVLChatModel"
+  ],
+  "auto_map": {
+    "AutoConfig": "configuration_internvl_chat.InternVLChatConfig",
+    "AutoModel": "modeling_internvl_chat.InternVLChatModel",
+    "AutoModelForCausalLM": "modeling_internvl_chat.InternVLChatModel"
+  },
+  "downsample_ratio": 0.5,
+  "dynamic_image_size": true,
+  "force_image_size": 448,
+  "llm_config": {
+    "_name_or_path": "internlm/internlm2_5-7b-chat",
+    "add_cross_attention": false,
+    "architectures": [
+      "InternLM2ForCausalLM"
+    ],
+    "attn_implementation": "flash_attention_2",
+    "auto_map": {
+      "AutoConfig": "configuration_internlm2.InternLM2Config",
+      "AutoModel": "modeling_internlm2.InternLM2ForCausalLM",
+      "AutoModelForCausalLM": "modeling_internlm2.InternLM2ForCausalLM"
+    },
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bias": false,
+    "bos_token_id": 1,
+    "chunk_size_feed_forward": 0,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": 2,
+    "exponential_decay_length_penalty": null,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "hidden_act": "silu",
+    "hidden_size": 4096,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "initializer_range": 0.02,
+    "intermediate_size": 14336,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "max_position_embeddings": 32768,
+    "min_length": 0,
+    "model_type": "internlm2",
+    "no_repeat_ngram_size": 0,
+    "num_attention_heads": 32,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_hidden_layers": 32,
+    "num_key_value_heads": 8,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "pad_token_id": 2,
+    "prefix": null,
+    "pretraining_tp": 1,
+    "problem_type": null,
+    "pruned_heads": {},
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "rms_norm_eps": 1e-05,
+    "rope_scaling": {
+      "factor": 2.0,
+      "type": "dynamic"
+    },
+    "rope_theta": 1000000,
+    "sep_token_id": null,
+    "suppress_tokens": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tf_legacy_loss": false,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": false,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": "bfloat16",
+    "torchscript": false,
+    "transformers_version": "4.39.0",
+    "typical_p": 1.0,
+    "use_bfloat16": true,
+    "use_cache": false,
+    "vocab_size": 92557
+  },
+  "max_dynamic_patch": 12,
+  "min_dynamic_patch": 1,
+  "model_type": "internvl_chat",
+  "pad2square": false,
+  "ps_version": "v2",
+  "select_layer": -1,
+  "template": "internlm2-chat",
+  "torch_dtype": "bfloat16",
+  "transformers_version": null,
+  "use_backbone_lora": 0,
+  "use_llm_lora": 0,
+  "use_thumbnail": true,
+  "vision_config": {
+    "_name_or_path": "",
+    "add_cross_attention": false,
+    "architectures": [
+      "InternVisionModel"
+    ],
+    "attention_dropout": 0.0,
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bos_token_id": null,
+    "chunk_size_feed_forward": 0,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "drop_path_rate": 0.1,
+    "dropout": 0.0,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": null,
+    "exponential_decay_length_penalty": null,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "hidden_act": "gelu",
+    "hidden_size": 1024,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "image_size": 448,
+    "initializer_factor": 1.0,
+    "initializer_range": 0.02,
+    "intermediate_size": 4096,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "layer_norm_eps": 1e-06,
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "min_length": 0,
+    "model_type": "intern_vit_6b",
+    "no_repeat_ngram_size": 0,
+    "norm_type": "layer_norm",
+    "num_attention_heads": 16,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_channels": 3,
+    "num_hidden_layers": 24,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "pad_token_id": null,
+    "patch_size": 14,
+    "prefix": null,
+    "problem_type": null,
+    "pruned_heads": {},
+    "qk_normalization": false,
+    "qkv_bias": true,
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "sep_token_id": null,
+    "suppress_tokens": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tf_legacy_loss": false,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": true,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": "bfloat16",
+    "torchscript": false,
+    "transformers_version": "4.39.0",
+    "typical_p": 1.0,
+    "use_bfloat16": true,
+    "use_flash_attn": true
+  }
+}

generation_config.json

@@ -0,0 +1,8 @@
+{
+  "_from_model_config": true,
+  "eos_token_id": [
+    92542,
+    92543
+  ],
+  "transformers_version": "4.39.0"
+}

model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:437cef79372990d54f873dea9f3ec904000fc4042db3aebe427f5d666da0f6b0
+size 4939977104

model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c3c315e5537cdff594bc905f77b1135a2382300fb5bcb1a6a32915f082b9a478
+size 4915914584

model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9078b00437074f05c2fc829ad8f441300b63e6d5866a2b4f6f044af7f0ae53d1
+size 4915914592

model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54096ec92708044d3f9650fb1d74bac94be3893af9336c6de2e20c5ba3f50b4e
+size 1379059688

model.safetensors.index.json

@@ -0,0 +1,580 @@
+{
+  "metadata": {
+    "total_size": 16150796288
+  },
+  "weight_map": {
+    "language_model.model.layers.0.attention.wo.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.0.attention.wqkv.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.0.attention_norm.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.0.feed_forward.w1.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.0.feed_forward.w2.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.0.feed_forward.w3.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.0.ffn_norm.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.1.attention.wo.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.1.attention.wqkv.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.1.attention_norm.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.1.feed_forward.w1.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.1.feed_forward.w2.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.1.feed_forward.w3.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.1.ffn_norm.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.10.attention.wo.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.10.attention.wqkv.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.10.attention_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.10.feed_forward.w1.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.10.feed_forward.w2.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.10.feed_forward.w3.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.10.ffn_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.11.attention.wo.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.11.attention.wqkv.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.11.attention_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.11.feed_forward.w1.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.11.feed_forward.w2.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.11.feed_forward.w3.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.11.ffn_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.12.attention.wo.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.12.attention.wqkv.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.12.attention_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.12.feed_forward.w1.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.12.feed_forward.w2.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.12.feed_forward.w3.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.12.ffn_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.13.attention.wo.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.13.attention.wqkv.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.13.attention_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.13.feed_forward.w1.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.13.feed_forward.w2.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.13.feed_forward.w3.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.13.ffn_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.14.attention.wo.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.14.attention.wqkv.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.14.attention_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.14.feed_forward.w1.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.14.feed_forward.w2.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.14.feed_forward.w3.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.14.ffn_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.15.attention.wo.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.15.attention.wqkv.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.15.attention_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.15.feed_forward.w1.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.15.feed_forward.w2.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.15.feed_forward.w3.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.15.ffn_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.16.attention.wo.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.16.attention.wqkv.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.16.attention_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.16.feed_forward.w1.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.16.feed_forward.w2.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.16.feed_forward.w3.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.16.ffn_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.17.attention.wo.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.17.attention.wqkv.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.17.attention_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.17.feed_forward.w1.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.17.feed_forward.w2.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.17.feed_forward.w3.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.17.ffn_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.18.attention.wo.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.18.attention.wqkv.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.18.attention_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.18.feed_forward.w1.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.18.feed_forward.w2.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.18.feed_forward.w3.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.18.ffn_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.19.attention.wo.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.19.attention.wqkv.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.19.attention_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.19.feed_forward.w1.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.19.feed_forward.w2.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.19.feed_forward.w3.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.19.ffn_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.2.attention.wo.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.2.attention.wqkv.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.2.attention_norm.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.2.feed_forward.w1.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.2.feed_forward.w2.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.2.feed_forward.w3.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.2.ffn_norm.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.20.attention.wo.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.20.attention.wqkv.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.20.attention_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.20.feed_forward.w1.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.20.feed_forward.w2.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.20.feed_forward.w3.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.20.ffn_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.21.attention.wo.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.21.attention.wqkv.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.21.attention_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.21.feed_forward.w1.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.21.feed_forward.w2.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.21.feed_forward.w3.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.21.ffn_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.22.attention.wo.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.22.attention.wqkv.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.22.attention_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.22.feed_forward.w1.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.22.feed_forward.w2.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.22.feed_forward.w3.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.22.ffn_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.23.attention.wo.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.23.attention.wqkv.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.23.attention_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.23.feed_forward.w1.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.23.feed_forward.w2.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.23.feed_forward.w3.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.23.ffn_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.24.attention.wo.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.24.attention.wqkv.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.24.attention_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.24.feed_forward.w1.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.24.feed_forward.w2.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.24.feed_forward.w3.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.24.ffn_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.25.attention.wo.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.25.attention.wqkv.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.25.attention_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.25.feed_forward.w1.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.25.feed_forward.w2.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.25.feed_forward.w3.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.25.ffn_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.26.attention.wo.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.26.attention.wqkv.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.26.attention_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.26.feed_forward.w1.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.26.feed_forward.w2.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.26.feed_forward.w3.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.26.ffn_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.27.attention.wo.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.27.attention.wqkv.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.27.attention_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.27.feed_forward.w1.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.27.feed_forward.w2.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.27.feed_forward.w3.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.27.ffn_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.28.attention.wo.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.28.attention.wqkv.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.28.attention_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.28.feed_forward.w1.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.28.feed_forward.w2.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.28.feed_forward.w3.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.28.ffn_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.29.attention.wo.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.29.attention.wqkv.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.29.attention_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.29.feed_forward.w1.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.29.feed_forward.w2.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.29.feed_forward.w3.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.29.ffn_norm.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.3.attention.wo.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.3.attention.wqkv.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.3.attention_norm.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.3.feed_forward.w1.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.3.feed_forward.w2.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.3.feed_forward.w3.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.3.ffn_norm.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.30.attention.wo.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.30.attention.wqkv.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.30.attention_norm.weight": "model-00004-of-00004.safetensors",
+    "language_model.model.layers.30.feed_forward.w1.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.30.feed_forward.w2.weight": "model-00004-of-00004.safetensors",
+    "language_model.model.layers.30.feed_forward.w3.weight": "model-00003-of-00004.safetensors",
+    "language_model.model.layers.30.ffn_norm.weight": "model-00004-of-00004.safetensors",
+    "language_model.model.layers.31.attention.wo.weight": "model-00004-of-00004.safetensors",
+    "language_model.model.layers.31.attention.wqkv.weight": "model-00004-of-00004.safetensors",
+    "language_model.model.layers.31.attention_norm.weight": "model-00004-of-00004.safetensors",
+    "language_model.model.layers.31.feed_forward.w1.weight": "model-00004-of-00004.safetensors",
+    "language_model.model.layers.31.feed_forward.w2.weight": "model-00004-of-00004.safetensors",
+    "language_model.model.layers.31.feed_forward.w3.weight": "model-00004-of-00004.safetensors",
+    "language_model.model.layers.31.ffn_norm.weight": "model-00004-of-00004.safetensors",
+    "language_model.model.layers.4.attention.wo.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.4.attention.wqkv.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.4.attention_norm.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.4.feed_forward.w1.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.4.feed_forward.w2.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.4.feed_forward.w3.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.4.ffn_norm.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.5.attention.wo.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.5.attention.wqkv.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.5.attention_norm.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.5.feed_forward.w1.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.5.feed_forward.w2.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.5.feed_forward.w3.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.5.ffn_norm.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.6.attention.wo.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.6.attention.wqkv.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.6.attention_norm.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.6.feed_forward.w1.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.6.feed_forward.w2.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.6.feed_forward.w3.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.6.ffn_norm.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.7.attention.wo.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.7.attention.wqkv.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.7.attention_norm.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.7.feed_forward.w1.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.7.feed_forward.w2.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.7.feed_forward.w3.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.7.ffn_norm.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.8.attention.wo.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.8.attention.wqkv.weight": "model-00001-of-00004.safetensors",
+    "language_model.model.layers.8.attention_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.8.feed_forward.w1.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.8.feed_forward.w2.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.8.feed_forward.w3.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.8.ffn_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.9.attention.wo.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.9.attention.wqkv.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.9.attention_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.9.feed_forward.w1.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.9.feed_forward.w2.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.9.feed_forward.w3.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.layers.9.ffn_norm.weight": "model-00002-of-00004.safetensors",
+    "language_model.model.norm.weight": "model-00004-of-00004.safetensors",
+    "language_model.model.tok_embeddings.weight": "model-00001-of-00004.safetensors",
+    "language_model.output.weight": "model-00004-of-00004.safetensors",
+    "mlp1.0.bias": "model-00004-of-00004.safetensors",
+    "mlp1.0.weight": "model-00004-of-00004.safetensors",
+    "mlp1.1.bias": "model-00004-of-00004.safetensors",
+    "mlp1.1.weight": "model-00004-of-00004.safetensors",
+    "mlp1.3.bias": "model-00004-of-00004.safetensors",
+    "mlp1.3.weight": "model-00004-of-00004.safetensors",
+    "vision_model.embeddings.class_embedding": "model-00001-of-00004.safetensors",
+    "vision_model.embeddings.patch_embedding.bias": "model-00001-of-00004.safetensors",
+    "vision_model.embeddings.patch_embedding.weight": "model-00001-of-00004.safetensors",
+    "vision_model.embeddings.position_embedding": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.0.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.0.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.0.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.0.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.0.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.0.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.0.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.0.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.0.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.0.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.0.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.0.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.0.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.0.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.1.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.1.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.1.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.1.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.1.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.1.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.1.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.1.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.1.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.1.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.1.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.1.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.1.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.1.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.10.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.10.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.10.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.10.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.10.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.10.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.10.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.10.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.10.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.10.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.10.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.10.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.10.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.10.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.11.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.11.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.11.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.11.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.11.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.11.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.11.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.11.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.11.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.11.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.11.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.11.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.11.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.11.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.12.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.12.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.12.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.12.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.12.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.12.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.12.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.12.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.12.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.12.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.12.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.12.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.12.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.12.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.13.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.13.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.13.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.13.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.13.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.13.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.13.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.13.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.13.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.13.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.13.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.13.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.13.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.13.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.14.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.14.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.14.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.14.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.14.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.14.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.14.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.14.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.14.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.14.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.14.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.14.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.14.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.14.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.15.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.15.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.15.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.15.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.15.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.15.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.15.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.15.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.15.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.15.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.15.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.15.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.15.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.15.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.16.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.16.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.16.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.16.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.16.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.16.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.16.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.16.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.16.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.16.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.16.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.16.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.16.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.16.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.17.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.17.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.17.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.17.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.17.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.17.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.17.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.17.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.17.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.17.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.17.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.17.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.17.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.17.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.18.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.18.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.18.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.18.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.18.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.18.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.18.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.18.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.18.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.18.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.18.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.18.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.18.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.18.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.19.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.19.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.19.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.19.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.19.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.19.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.19.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.19.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.19.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.19.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.19.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.19.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.19.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.19.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.2.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.2.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.2.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.2.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.2.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.2.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.2.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.2.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.2.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.2.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.2.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.2.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.2.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.2.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.20.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.20.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.20.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.20.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.20.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.20.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.20.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.20.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.20.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.20.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.20.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.20.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.20.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.20.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.21.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.21.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.21.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.21.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.21.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.21.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.21.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.21.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.21.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.21.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.21.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.21.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.21.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.21.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.22.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.22.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.22.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.22.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.22.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.22.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.22.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.22.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.22.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.22.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.22.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.22.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.22.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.22.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.23.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.23.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.23.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.23.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.23.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.23.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.23.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.23.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.23.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.23.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.23.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.23.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.23.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.23.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.3.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.3.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.3.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.3.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.3.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.3.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.3.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.3.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.3.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.3.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.3.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.3.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.3.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.3.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.4.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.4.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.4.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.4.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.4.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.4.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.4.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.4.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.4.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.4.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.4.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.4.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.4.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.4.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.5.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.5.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.5.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.5.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.5.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.5.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.5.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.5.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.5.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.5.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.5.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.5.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.5.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.5.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.6.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.6.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.6.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.6.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.6.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.6.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.6.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.6.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.6.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.6.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.6.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.6.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.6.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.6.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.7.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.7.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.7.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.7.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.7.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.7.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.7.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.7.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.7.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.7.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.7.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.7.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.7.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.7.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.8.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.8.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.8.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.8.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.8.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.8.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.8.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.8.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.8.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.8.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.8.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.8.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.8.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.8.norm2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.9.attn.proj.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.9.attn.proj.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.9.attn.qkv.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.9.attn.qkv.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.9.ls1": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.9.ls2": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.9.mlp.fc1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.9.mlp.fc1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.9.mlp.fc2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.9.mlp.fc2.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.9.norm1.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.9.norm1.weight": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.9.norm2.bias": "model-00001-of-00004.safetensors",
+    "vision_model.encoder.layers.9.norm2.weight": "model-00001-of-00004.safetensors"
+  }
+}

special_tokens_map.json

@@ -0,0 +1,75 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|action_start|>",
+    "<|action_end|>",
+    "<|interpreter|>",
+    "<|plugin|>",
+    "<img>",
+    "</img>",
+    "<IMG_CONTEXT>",
+    "<quad>",
+    "</quad>",
+    "<ref>",
+    "</ref>",
+    "<box>",
+    "</box>",
+    {
+      "content": "<|python|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|/python|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|execution|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|/execution|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenization_internlm2.py

@@ -0,0 +1,235 @@
+# Copyright (c) The InternLM team and The HuggingFace Inc. team. All rights reserved.
+#
+# This code is based on transformers/src/transformers/models/llama/tokenization_llama.py
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Tokenization classes for InternLM."""
+import os
+from shutil import copyfile
+from typing import Any, Dict, List, Optional, Tuple
+
+import sentencepiece as spm
+from transformers.tokenization_utils import PreTrainedTokenizer
+from transformers.utils import logging
+
+logger = logging.get_logger(__name__)
+
+VOCAB_FILES_NAMES = {'vocab_file': './tokenizer.model'}
+
+PRETRAINED_VOCAB_FILES_MAP = {}
+
+
+# Modified from transformers.model.llama.tokenization_llama.LlamaTokenizer
+class InternLM2Tokenizer(PreTrainedTokenizer):
+    """
+    Construct a InternLM2 tokenizer. Based on byte-level Byte-Pair-Encoding.
+
+    Args:
+        vocab_file (`str`):
+            Path to the vocabulary file.
+    """
+
+    vocab_files_names = VOCAB_FILES_NAMES
+    pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP
+    model_input_names = ['input_ids', 'attention_mask']
+    _auto_class = 'AutoTokenizer'
+
+    def __init__(
+        self,
+        vocab_file,
+        unk_token='<unk>',
+        bos_token='<s>',
+        eos_token='</s>',
+        pad_token='</s>',
+        sp_model_kwargs: Optional[Dict[str, Any]] = None,
+        add_bos_token=True,
+        add_eos_token=False,
+        decode_with_prefix_space=False,
+        clean_up_tokenization_spaces=False,
+        **kwargs,
+    ):
+        self.sp_model_kwargs = {} if sp_model_kwargs is None else sp_model_kwargs
+        self.vocab_file = vocab_file
+        self.add_bos_token = add_bos_token
+        self.add_eos_token = add_eos_token
+        self.decode_with_prefix_space = decode_with_prefix_space
+        self.sp_model = spm.SentencePieceProcessor(**self.sp_model_kwargs)
+        self.sp_model.Load(vocab_file)
+        self._no_prefix_space_tokens = None
+        super().__init__(
+            bos_token=bos_token,
+            eos_token=eos_token,
+            unk_token=unk_token,
+            pad_token=pad_token,
+            clean_up_tokenization_spaces=clean_up_tokenization_spaces,
+            **kwargs,
+        )
+
+    @property
+    def no_prefix_space_tokens(self):
+        if self._no_prefix_space_tokens is None:
+            vocab = self.convert_ids_to_tokens(list(range(self.vocab_size)))
+            self._no_prefix_space_tokens = {i for i, tok in enumerate(vocab) if not tok.startswith('▁')}
+        return self._no_prefix_space_tokens
+
+    @property
+    def vocab_size(self):
+        """Returns vocab size"""
+        return self.sp_model.get_piece_size()
+
+    @property
+    def bos_token_id(self) -> Optional[int]:
+        return self.sp_model.bos_id()
+
+    @property
+    def eos_token_id(self) -> Optional[int]:
+        return self.sp_model.eos_id()
+
+    def get_vocab(self):
+        """Returns vocab as a dict"""
+        vocab = {self.convert_ids_to_tokens(i): i for i in range(self.vocab_size)}
+        vocab.update(self.added_tokens_encoder)
+        return vocab
+
+    def _tokenize(self, text):
+        """Returns a tokenized string."""
+        return self.sp_model.encode(text, out_type=str)
+
+    def _convert_token_to_id(self, token):
+        """Converts a token (str) in an id using the vocab."""
+        return self.sp_model.piece_to_id(token)
+
+    def _convert_id_to_token(self, index):
+        """Converts an index (integer) in a token (str) using the vocab."""
+        token = self.sp_model.IdToPiece(index)
+        return token
+
+    def _maybe_add_prefix_space(self, tokens, decoded):
+        if tokens and tokens[0] not in self.no_prefix_space_tokens:
+            return ' ' + decoded
+        else:
+            return decoded
+
+    def convert_tokens_to_string(self, tokens):
+        """Converts a sequence of tokens (string) in a single string."""
+        current_sub_tokens = []
+        out_string = ''
+        prev_is_special = False
+        for token in tokens:
+            # make sure that special tokens are not decoded using sentencepiece model
+            if token in self.all_special_tokens:
+                if not prev_is_special:
+                    out_string += ' '
+                out_string += self.sp_model.decode(current_sub_tokens) + token
+                prev_is_special = True
+                current_sub_tokens = []
+            else:
+                current_sub_tokens.append(token)
+                prev_is_special = False
+        out_string += self.sp_model.decode(current_sub_tokens)
+        out_string = self.clean_up_tokenization(out_string)
+        out_string = self._maybe_add_prefix_space(tokens=tokens, decoded=out_string)
+        return out_string[1:]
+
+    def save_vocabulary(self, save_directory, filename_prefix: Optional[str] = None) -> Tuple[str]:
+        """
+        Save the vocabulary and special tokens file to a directory.
+
+        Args:
+            save_directory (`str`):
+                The directory in which to save the vocabulary.
+
+        Returns:
+            `Tuple(str)`: Paths to the files saved.
+        """
+        if not os.path.isdir(save_directory):
+            logger.error(f'Vocabulary path ({save_directory}) should be a directory')
+            return
+        out_vocab_file = os.path.join(
+            save_directory, (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file']
+        )
+
+        if os.path.abspath(self.vocab_file) != os.path.abspath(out_vocab_file) and os.path.isfile(self.vocab_file):
+            copyfile(self.vocab_file, out_vocab_file)
+        elif not os.path.isfile(self.vocab_file):
+            with open(out_vocab_file, 'wb') as fi:
+                content_spiece_model = self.sp_model.serialized_model_proto()
+                fi.write(content_spiece_model)
+
+        return (out_vocab_file,)
+
+    def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
+        if self.add_bos_token:
+            bos_token_ids = [self.bos_token_id]
+        else:
+            bos_token_ids = []
+
+        output = bos_token_ids + token_ids_0
+
+        if token_ids_1 is not None:
+            output = output + token_ids_1
+
+        if self.add_eos_token:
+            output = output + [self.eos_token_id]
+
+        return output
+
+    def get_special_tokens_mask(
+        self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None, already_has_special_tokens: bool = False
+    ) -> List[int]:
+        """
+        Retrieve sequence ids from a token list that has no special tokens added. This method is called when adding
+        special tokens using the tokenizer `prepare_for_model` method.
+
+        Args:
+            token_ids_0 (`List[int]`):
+                List of IDs.
+            token_ids_1 (`List[int]`, *optional*):
+                Optional second list of IDs for sequence pairs.
+            already_has_special_tokens (`bool`, *optional*, defaults to `False`):
+                Whether or not the token list is already formatted with special tokens for the model.
+
+        Returns:
+            `List[int]`: A list of integers in the range [0, 1]: 1 for a special token, 0 for a sequence token.
+        """
+        if already_has_special_tokens:
+            return super().get_special_tokens_mask(
+                token_ids_0=token_ids_0, token_ids_1=token_ids_1, already_has_special_tokens=True
+            )
+
+        if token_ids_1 is None:
+            return [1] + ([0] * len(token_ids_0)) + [1]
+        return [1] + ([0] * len(token_ids_0)) + [1, 1] + ([0] * len(token_ids_1)) + [1]
+
+    def create_token_type_ids_from_sequences(
+        self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
+    ) -> List[int]:
+        """
+        Create a mask from the two sequences passed to be used in a sequence-pair classification task. T5 does not make
+        use of token type ids, therefore a list of zeros is returned.
+
+        Args:
+            token_ids_0 (`List[int]`):
+                List of IDs.
+            token_ids_1 (`List[int]`, *optional*):
+                Optional second list of IDs for sequence pairs.
+
+        Returns:
+            `List[int]`: List of zeros.
+        """
+        eos = [self.eos_token_id]
+
+        if token_ids_1 is None:
+            return len(token_ids_0 + eos) * [0]
+        return len(token_ids_0 + eos + token_ids_1 + eos) * [0]

tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f868398fc4e05ee1e8aeba95ddf18ddcc45b8bce55d5093bead5bbf80429b48b
+size 1477754

tokenizer_config.json

@@ -0,0 +1,215 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92538": {
+      "content": "<|plugin|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92539": {
+      "content": "<|interpreter|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92540": {
+      "content": "<|action_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92541": {
+      "content": "<|action_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92542": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92543": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92544": {
+      "content": "<img>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92545": {
+      "content": "</img>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92546": {
+      "content": "<IMG_CONTEXT>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92547": {
+      "content": "<quad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92548": {
+      "content": "</quad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92549": {
+      "content": "<ref>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92550": {
+      "content": "</ref>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92551": {
+      "content": "<box>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92552": {
+      "content": "</box>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92553": {
+      "content": "<|python|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92554": {
+      "content": "<|/python|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92555": {
+      "content": "<|execution|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "92556": {
+      "content": "<|/execution|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|action_start|>",
+    "<|action_end|>",
+    "<|interpreter|>",
+    "<|plugin|>",
+    "<img>",
+    "</img>",
+    "<IMG_CONTEXT>",
+    "<quad>",
+    "</quad>",
+    "<ref>",
+    "</ref>",
+    "<box>",
+    "</box>",
+    "<|python|>",
+    "<|/python|>",
+    "<|execution|>",
+    "<|/execution|>"
+  ],
+  "auto_map": {
+    "AutoTokenizer": [
+      "tokenization_internlm2.InternLM2Tokenizer",
+      null
+    ]
+  },
+  "bos_token": "<s>",
+  "chat_template": "{{ bos_token }}{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "model_max_length": 4096,
+  "pad_token": "</s>",
+  "tokenizer_class": "InternLM2Tokenizer",
+  "unk_token": "<unk>"
+}