Upload folder using huggingface_hub

README.md

@@ -0,0 +1,79 @@
+---
+license: other
+datasets:
+- vietgpt/wikipedia_vi
+- wikipedia
+- pg19
+- mc4
+language:
+- vi
+- en
+---
+
+We employed [SentencePiece](https://github.com/google/sentencepiece) to retrain a Vietnamese tokenizer with a vocabulary size of 20K. No Vietnamese word segmentation was used. We then merged this vocabulary with the original one of Llama2, removing duplicate tokens.
+The new tokenizer significantly improves when encoding Vietnamese text, reducing the number of tokens by 50% compared to ChatGPT and approximately 70% compared to the original Llama2.
+
+We conducted a single-epoch continual pretraining, also known as incremental pretraining, using the Llama2-chat 7B model on a mixed dataset totaling 40.5 GB, comprised of:
+- 19 GB [NewsCorpus](https://github.com/binhvq/news-corpus)
+- 1.1 GB Vietnamese Wikipedia
+- 1.6 GB [Vietnamese books](https://www.kaggle.com/datasets/iambestfeeder/10000-vietnamese-books)
+- 4.5 GB Vietnamese legal documents (crawled from thuvienphapluat and processed by ourselves)
+- 2.1 GB Vietnamese legal text (from [C4-vi](https://huggingface.co/datasets/c4))
+- 1.1 GB English Books (sub-sampled from [pg19](https://huggingface.co/datasets/pg19))
+- 1.1 GB English Wikipedia (sub-sampled from 20220301.en wikipedia)
+- 10 GB English Text (sub-sampled from [C4-en](https://huggingface.co/datasets/c4))
+
+We trained the model on a DGX A100 system, utilizing four GPU A100 in 10 days (about 1000 GPU hours). 
+
+Hyperparameters are set as follows:
+- Training Regime: BFloat16 mixed precision
+- Lora Config: 
+  
+  ```
+  {
+    "base_model_name_or_path": "meta-llama/Llama-2-7b-chat-hf",
+    "bias": "none",
+    "enable_lora": null,
+    "fan_in_fan_out": false,
+    "inference_mode": true,
+    "lora_alpha": 32.0,
+    "lora_dropout": 0.05,
+    "merge_weights": false,
+    "modules_to_save": [
+      "embed_tokens",
+      "lm_head"
+    ],
+    "peft_type": "LORA",
+    "r": 8,
+    "target_modules": [
+      "q_proj",
+      "v_proj",
+      "k_proj",
+      "o_proj",
+      "gate_proj",
+      "down_proj",
+      "up_proj"
+    ],
+    "task_type": "CAUSAL_LM"
+  }
+  
+  ```
+
+We also provide the [LoRA part](https://huggingface.co/bkai-foundation-models/vietnamese-llama2-7b-40GB/tree/main/pt_lora_model) so that you can integrate it with the original Llama2-chat-7b by yourself.
+
+Please note that **this model requires further supervised fine-tuning (SFT)** to be used in practice!
+
+Usage and other considerations: We refer to the [Llama 2](https://github.com/facebookresearch/llama)
+
+Training loss:
+<img src="figure/training_loss.png" alt="Training Loss Curve"/>
+
+**Disclaimer**
+
+This project is built upon Meta's Llama-2 model. It is essential to strictly adhere to the open-source license agreement of Llama-2 when using this model. If you incorporate third-party code, please ensure compliance with the relevant open-source license agreements.
+It's important to note that the content generated by the model may be influenced by various factors, such as calculation methods, random elements, and potential inaccuracies in quantification. Consequently, this project does not offer any guarantees regarding the accuracy of the model's outputs, and it disclaims any responsibility for consequences resulting from the use of the model's resources and its output.
+For those employing the models from this project for commercial purposes, developers must adhere to local laws and regulations to ensure the compliance of the model's output content. This project is not accountable for any products or services derived from such usage.
+
+**Acknowledgments**
+
+We extend our gratitude to PHPC - Phenikaa University and NVIDIA for their generous provision of computing resources for model training. Our appreciation also goes out to binhvq and the other authors for their diligent efforts in collecting and preparing the Vietnamese text corpus.

config.json

@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "bkai-foundation-models/vietnamese-llama2-7b-40GB",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "max_position_embeddings": 4096,
+  "model_type": "llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 32,
+  "pad_token_id": 0,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float16",
+  "transformers_version": "4.31.0",
+  "use_cache": true,
+  "vocab_size": 46303
+}

generation_config.json

@@ -0,0 +1,17 @@
+{
+  "bos_token_id": 1,
+  "do_sample": true,
+  "eos_token_id": 2,
+  "max_length": 4096,
+  "max_memory": {
+    "cpu": 329543315456
+  },
+  "no_split_module_classes": [
+    "LlamaDecoderLayer"
+  ],
+  "pad_token_id": 0,
+  "special_dtypes": {},
+  "temperature": 0.6,
+  "top_p": 0.9,
+  "transformers_version": "4.31.0"
+}

pt_lora_model/adapter_config.json

@@ -0,0 +1,26 @@
+{
+  "base_model_name_or_path": "meta-llama/Llama-2-7b-chat-hf",
+  "bias": "none",
+  "enable_lora": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "lora_alpha": 32.0,
+  "lora_dropout": 0.05,
+  "merge_weights": false,
+  "modules_to_save": [
+    "embed_tokens",
+    "lm_head"
+  ],
+  "peft_type": "LORA",
+  "r": 8,
+  "target_modules": [
+    "q_proj",
+    "v_proj",
+    "k_proj",
+    "o_proj",
+    "gate_proj",
+    "down_proj",
+    "up_proj"
+  ],
+  "task_type": "CAUSAL_LM"
+}

pt_lora_model/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8948e68d4459db8264be5b1f04c3ec06a926c1027b0db862d93462994eb04f2e
+size 798676157

pt_lora_model/special_tokens_map.json

@@ -0,0 +1,23 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  }
+}

pt_lora_model/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d92fd32e07ffd47cf903b4351e6ebc77f62afbc94bf895a2bfadfcd39cbf61e2
+size 728273

pt_lora_model/tokenizer_config.json

@@ -0,0 +1,35 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "bos_token": {
+    "__type": "AddedToken",
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "clean_up_tokenization_spaces": false,
+  "eos_token": {
+    "__type": "AddedToken",
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "legacy": false,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": null,
+  "sp_model_kwargs": {},
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": {
+    "__type": "AddedToken",
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "use_fast": true
+}

pytorch_model.bin.index.json

@@ -0,0 +1,330 @@
+{
+  "metadata": {
+    "total_size": 13711179776
+  },
+  "weight_map": {
+    "lm_head.weight": "pytorch_model_00008-of-00010.bin",
+    "model.embed_tokens.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.0.input_layernorm.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.0.mlp.down_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.0.mlp.gate_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.0.mlp.up_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.0.post_attention_layernorm.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.0.self_attn.k_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.0.self_attn.o_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.0.self_attn.q_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.0.self_attn.rotary_emb.inv_freq": "pytorch_model_00001-of-00010.bin",
+    "model.layers.0.self_attn.v_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.1.input_layernorm.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.1.mlp.down_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.1.mlp.gate_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.1.mlp.up_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.1.post_attention_layernorm.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.1.self_attn.k_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.1.self_attn.o_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.1.self_attn.q_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.1.self_attn.rotary_emb.inv_freq": "pytorch_model_00001-of-00010.bin",
+    "model.layers.1.self_attn.v_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.10.input_layernorm.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.10.mlp.down_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.10.mlp.gate_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.10.mlp.up_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.10.post_attention_layernorm.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.10.self_attn.k_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.10.self_attn.o_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.10.self_attn.q_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.10.self_attn.rotary_emb.inv_freq": "pytorch_model_00001-of-00010.bin",
+    "model.layers.10.self_attn.v_proj.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.11.input_layernorm.weight": "pytorch_model_00001-of-00010.bin",
+    "model.layers.11.mlp.down_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.11.mlp.gate_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.11.mlp.up_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.11.post_attention_layernorm.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.11.self_attn.k_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.11.self_attn.o_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.11.self_attn.q_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.11.self_attn.rotary_emb.inv_freq": "pytorch_model_00002-of-00010.bin",
+    "model.layers.11.self_attn.v_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.12.input_layernorm.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.12.mlp.down_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.12.mlp.gate_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.12.mlp.up_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.12.post_attention_layernorm.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.12.self_attn.k_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.12.self_attn.o_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.12.self_attn.q_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.12.self_attn.rotary_emb.inv_freq": "pytorch_model_00002-of-00010.bin",
+    "model.layers.12.self_attn.v_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.13.input_layernorm.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.13.mlp.down_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.13.mlp.gate_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.13.mlp.up_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.13.post_attention_layernorm.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.13.self_attn.k_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.13.self_attn.o_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.13.self_attn.q_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.13.self_attn.rotary_emb.inv_freq": "pytorch_model_00002-of-00010.bin",
+    "model.layers.13.self_attn.v_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.14.input_layernorm.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.14.mlp.down_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.14.mlp.gate_proj.weight": "pytorch_model_00002-of-00010.bin",
+    "model.layers.14.mlp.up_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.14.post_attention_layernorm.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.14.self_attn.k_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.14.self_attn.o_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.14.self_attn.q_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.14.self_attn.rotary_emb.inv_freq": "pytorch_model_00003-of-00010.bin",
+    "model.layers.14.self_attn.v_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.15.input_layernorm.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.15.mlp.down_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.15.mlp.gate_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.15.mlp.up_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.15.post_attention_layernorm.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.15.self_attn.k_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.15.self_attn.o_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.15.self_attn.q_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.15.self_attn.rotary_emb.inv_freq": "pytorch_model_00003-of-00010.bin",
+    "model.layers.15.self_attn.v_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.16.input_layernorm.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.16.mlp.down_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.16.mlp.gate_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.16.mlp.up_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.16.post_attention_layernorm.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.16.self_attn.k_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.16.self_attn.o_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.16.self_attn.q_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.16.self_attn.rotary_emb.inv_freq": "pytorch_model_00003-of-00010.bin",
+    "model.layers.16.self_attn.v_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.17.input_layernorm.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.17.mlp.down_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.17.mlp.gate_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.17.mlp.up_proj.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.17.post_attention_layernorm.weight": "pytorch_model_00003-of-00010.bin",
+    "model.layers.17.self_attn.k_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.17.self_attn.o_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.17.self_attn.q_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.17.self_attn.rotary_emb.inv_freq": "pytorch_model_00004-of-00010.bin",
+    "model.layers.17.self_attn.v_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.18.input_layernorm.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.18.mlp.down_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.18.mlp.gate_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.18.mlp.up_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.18.post_attention_layernorm.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.18.self_attn.k_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.18.self_attn.o_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.18.self_attn.q_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.18.self_attn.rotary_emb.inv_freq": "pytorch_model_00004-of-00010.bin",
+    "model.layers.18.self_attn.v_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.19.input_layernorm.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.19.mlp.down_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.19.mlp.gate_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.19.mlp.up_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.19.post_attention_layernorm.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.19.self_attn.k_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.19.self_attn.o_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.19.self_attn.q_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.19.self_attn.rotary_emb.inv_freq": "pytorch_model_00004-of-00010.bin",
+    "model.layers.19.self_attn.v_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.2.input_layernorm.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.2.mlp.down_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.2.mlp.gate_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.2.mlp.up_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.2.post_attention_layernorm.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.2.self_attn.k_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.2.self_attn.o_proj.weight": "pytorch_model_00004-of-00010.bin",
+    "model.layers.2.self_attn.q_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.2.self_attn.rotary_emb.inv_freq": "pytorch_model_00005-of-00010.bin",
+    "model.layers.2.self_attn.v_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.20.input_layernorm.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.20.mlp.down_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.20.mlp.gate_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.20.mlp.up_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.20.post_attention_layernorm.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.20.self_attn.k_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.20.self_attn.o_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.20.self_attn.q_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.20.self_attn.rotary_emb.inv_freq": "pytorch_model_00005-of-00010.bin",
+    "model.layers.20.self_attn.v_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.21.input_layernorm.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.21.mlp.down_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.21.mlp.gate_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.21.mlp.up_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.21.post_attention_layernorm.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.21.self_attn.k_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.21.self_attn.o_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.21.self_attn.q_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.21.self_attn.rotary_emb.inv_freq": "pytorch_model_00005-of-00010.bin",
+    "model.layers.21.self_attn.v_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.22.input_layernorm.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.22.mlp.down_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.22.mlp.gate_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.22.mlp.up_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.22.post_attention_layernorm.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.22.self_attn.k_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.22.self_attn.o_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.22.self_attn.q_proj.weight": "pytorch_model_00005-of-00010.bin",
+    "model.layers.22.self_attn.rotary_emb.inv_freq": "pytorch_model_00005-of-00010.bin",
+    "model.layers.22.self_attn.v_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.23.input_layernorm.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.23.mlp.down_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.23.mlp.gate_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.23.mlp.up_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.23.post_attention_layernorm.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.23.self_attn.k_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.23.self_attn.o_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.23.self_attn.q_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.23.self_attn.rotary_emb.inv_freq": "pytorch_model_00006-of-00010.bin",
+    "model.layers.23.self_attn.v_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.24.input_layernorm.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.24.mlp.down_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.24.mlp.gate_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.24.mlp.up_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.24.post_attention_layernorm.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.24.self_attn.k_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.24.self_attn.o_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.24.self_attn.q_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.24.self_attn.rotary_emb.inv_freq": "pytorch_model_00008-of-00010.bin",
+    "model.layers.24.self_attn.v_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.25.input_layernorm.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.25.mlp.down_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.25.mlp.gate_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.25.mlp.up_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.25.post_attention_layernorm.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.25.self_attn.k_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.25.self_attn.o_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.25.self_attn.q_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.25.self_attn.rotary_emb.inv_freq": "pytorch_model_00009-of-00010.bin",
+    "model.layers.25.self_attn.v_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.26.input_layernorm.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.26.mlp.down_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.26.mlp.gate_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.26.mlp.up_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.26.post_attention_layernorm.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.26.self_attn.k_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.26.self_attn.o_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.26.self_attn.q_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.26.self_attn.rotary_emb.inv_freq": "pytorch_model_00009-of-00010.bin",
+    "model.layers.26.self_attn.v_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.27.input_layernorm.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.27.mlp.down_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.27.mlp.gate_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.27.mlp.up_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.27.post_attention_layernorm.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.27.self_attn.k_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.27.self_attn.o_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.27.self_attn.q_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.27.self_attn.rotary_emb.inv_freq": "pytorch_model_00009-of-00010.bin",
+    "model.layers.27.self_attn.v_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.28.input_layernorm.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.28.mlp.down_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.28.mlp.gate_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.28.mlp.up_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.28.post_attention_layernorm.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.28.self_attn.k_proj.weight": "pytorch_model_00009-of-00010.bin",
+    "model.layers.28.self_attn.o_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.28.self_attn.q_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.28.self_attn.rotary_emb.inv_freq": "pytorch_model_00010-of-00010.bin",
+    "model.layers.28.self_attn.v_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.29.input_layernorm.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.29.mlp.down_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.29.mlp.gate_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.29.mlp.up_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.29.post_attention_layernorm.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.29.self_attn.k_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.29.self_attn.o_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.29.self_attn.q_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.29.self_attn.rotary_emb.inv_freq": "pytorch_model_00010-of-00010.bin",
+    "model.layers.29.self_attn.v_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.3.input_layernorm.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.3.mlp.down_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.3.mlp.gate_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.3.mlp.up_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.3.post_attention_layernorm.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.3.self_attn.k_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.3.self_attn.o_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.3.self_attn.q_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.3.self_attn.rotary_emb.inv_freq": "pytorch_model_00006-of-00010.bin",
+    "model.layers.3.self_attn.v_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.30.input_layernorm.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.30.mlp.down_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.30.mlp.gate_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.30.mlp.up_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.30.post_attention_layernorm.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.30.self_attn.k_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.30.self_attn.o_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.30.self_attn.q_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.30.self_attn.rotary_emb.inv_freq": "pytorch_model_00010-of-00010.bin",
+    "model.layers.30.self_attn.v_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.31.input_layernorm.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.31.mlp.down_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.31.mlp.gate_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.31.mlp.up_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.31.post_attention_layernorm.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.31.self_attn.k_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.31.self_attn.o_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.31.self_attn.q_proj.weight": "pytorch_model_00010-of-00010.bin",
+    "model.layers.31.self_attn.rotary_emb.inv_freq": "pytorch_model_00011-of-00010.bin",
+    "model.layers.31.self_attn.v_proj.weight": "pytorch_model_00011-of-00010.bin",
+    "model.layers.4.input_layernorm.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.4.mlp.down_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.4.mlp.gate_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.4.mlp.up_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.4.post_attention_layernorm.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.4.self_attn.k_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.4.self_attn.o_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.4.self_attn.q_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.4.self_attn.rotary_emb.inv_freq": "pytorch_model_00006-of-00010.bin",
+    "model.layers.4.self_attn.v_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.5.input_layernorm.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.5.mlp.down_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.5.mlp.gate_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.5.mlp.up_proj.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.5.post_attention_layernorm.weight": "pytorch_model_00006-of-00010.bin",
+    "model.layers.5.self_attn.k_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.5.self_attn.o_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.5.self_attn.q_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.5.self_attn.rotary_emb.inv_freq": "pytorch_model_00007-of-00010.bin",
+    "model.layers.5.self_attn.v_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.6.input_layernorm.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.6.mlp.down_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.6.mlp.gate_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.6.mlp.up_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.6.post_attention_layernorm.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.6.self_attn.k_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.6.self_attn.o_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.6.self_attn.q_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.6.self_attn.rotary_emb.inv_freq": "pytorch_model_00007-of-00010.bin",
+    "model.layers.6.self_attn.v_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.7.input_layernorm.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.7.mlp.down_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.7.mlp.gate_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.7.mlp.up_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.7.post_attention_layernorm.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.7.self_attn.k_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.7.self_attn.o_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.7.self_attn.q_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.7.self_attn.rotary_emb.inv_freq": "pytorch_model_00007-of-00010.bin",
+    "model.layers.7.self_attn.v_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.8.input_layernorm.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.8.mlp.down_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.8.mlp.gate_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.8.mlp.up_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.8.post_attention_layernorm.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.8.self_attn.k_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.8.self_attn.o_proj.weight": "pytorch_model_00007-of-00010.bin",
+    "model.layers.8.self_attn.q_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.8.self_attn.rotary_emb.inv_freq": "pytorch_model_00008-of-00010.bin",
+    "model.layers.8.self_attn.v_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.9.input_layernorm.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.9.mlp.down_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.9.mlp.gate_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.9.mlp.up_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.9.post_attention_layernorm.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.9.self_attn.k_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.9.self_attn.o_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.9.self_attn.q_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.layers.9.self_attn.rotary_emb.inv_freq": "pytorch_model_00008-of-00010.bin",
+    "model.layers.9.self_attn.v_proj.weight": "pytorch_model_00008-of-00010.bin",
+    "model.norm.weight": "pytorch_model_00011-of-00010.bin"
+  }
+}

pytorch_model_00001-of-00010.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3108b9b6b0488d26208c5e2ce709e36ba131f2bce503a4c0a259b32aa6c4db69
+size 1593634410

pytorch_model_00002-of-00010.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0db711a538b166a34e28871055029499d99b8e09b9d5dc803210ccc258f0d8a
+size 1394667178

pytorch_model_00003-of-00010.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e0c56db6ec570d42ec629cda8f39b09b595d8115a69b0ad3b7b2fc5a15b8c3a
+size 1304497834

pytorch_model_00004-of-00010.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2e1f20fbe0e0162937d7585860ce4b1ecd5ec52cd848df1cbf1843cca12d89be
+size 1281420970

pytorch_model_00005-of-00010.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:90b92006422b1740ba980c05514908c2e614bb53178912f47df68e73675fa00a
+size 1247866794

pytorch_model_00006-of-00010.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b686e5f2e9ac67bd496baef33891398db5eed9987f82c6d6c7d740c7a4986c6
+size 1338044010

pytorch_model_00007-of-00010.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f2f2cf057d9a14a0cfe8c3132dc464726d540fb1336979420a7f34259d95a0ac
+size 1281420906

pytorch_model_00008-of-00010.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bbcf6bbd639fba2d44e178d1e6c64d17e46cdb7db05244fa16c7ea56b2a32cff
+size 1706880618

pytorch_model_00009-of-00010.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:388902ce76b9a1ad834c80cbca17c6dfc3a137be6de086d298e5022977bf8068
+size 1247874730

pytorch_model_00010-of-00010.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:56572ccb422f93fc9b5ffed02e1a755e4d1dcb72bd7d4e02a117bd5f8e8983ba
+size 1281420970

pytorch_model_00011-of-00010.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b6cfcbb35165452a1f358eeead96c05639838f0da274c69a4ab9984b67613b9e
+size 33564903

special_tokens_map.json

@@ -0,0 +1,23 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer_config.json

@@ -0,0 +1,35 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "bos_token": {
+    "__type": "AddedToken",
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "clean_up_tokenization_spaces": false,
+  "eos_token": {
+    "__type": "AddedToken",
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "legacy": false,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": null,
+  "sp_model_kwargs": {},
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": {
+    "__type": "AddedToken",
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "use_fast": true
+}