Upload Salesforce/codegen2-7B ctranslate fp16 weights

README.md

@@ -0,0 +1,130 @@
+
+---
+tags:
+- fauxpilot
+- gpt-j
+- float16
+
+license: apache-2.0
+---
+# Conversion for FauxPilot, Codegen-2 as GPT-J
+It feels like GPT-J, acts like any other GPT-J, but its Codegen-2 weights under the hood.
+
+Converted on 2023-05-22 using
+```
+python /home/michael/fauxpilot/converter/codegen_gptj_convert.py --code_model Salesforce/codegen2-7B /home/michael/tmp-codegen2-7B-gptj
+```
+
+# Licence and other remarks:
+Licence conditions are intended to be idential to original huggingface repo.
+
+# Original description
+see https://huggingface.co/'Salesforce/codegen2-7B'
+
+    
+
+# CodeGen2 (CodeGen2-16B)
+
+## Model description
+
+[CodeGen2](https://github.com/salesforce/CodeGen2) is a family of autoregressive language models for **program synthesis**, introduced in the paper:
+
+[CodeGen2: Lessons for Training LLMs on Programming and Natural Languages](https://arxiv.org/abs/2305.02309) by Erik Nijkamp\*, Hiroaki Hayashi\*, Caiming Xiong, Silvio Savarese, Yingbo Zhou.
+
+Unlike the original CodeGen model family (i.e., CodeGen1), CodeGen2 is capable of infilling, and supports more programming languages.
+
+Four model sizes are released: `1B`, `3.7B`, `7B`, `16B`.
+
+## How to use
+
+This model can be easily loaded using the `AutoModelForCausalLM` functionality.
+
+### Causal sampling
+
+For regular causal sampling, simply generate completions given the context:
+```python
+from transformers import AutoTokenizer, AutoModelForCausalLM
+tokenizer = AutoTokenizer.from_pretrained("Salesforce/codegen2-16B")
+model = AutoModelForCausalLM.from_pretrained("Salesforce/codegen2-16B", trust_remote_code=True, revision="main")
+
+text = "def hello_world():"
+input_ids = tokenizer(text, return_tensors="pt").input_ids
+generated_ids = model.generate(input_ids, max_length=128)
+print(tokenizer.decode(generated_ids[0], skip_special_tokens=True))
+```
+
+### Infill sampling
+
+For **infill** sampling, we introduce three new special token types:
+
+* `<mask_N>`: N-th span to be masked. In practice, use `<mask_1>` to where you want to sample infill.
+* `<sep>`: Seperator token between the suffix and the infilled sample. See below.
+* `<eom>`: "End-Of-Mask" token that model will output at the end of infilling. You may use this token to truncate the output.
+
+For example, if we want to generate infill for the following cursor position of a function:
+```python
+def hello_world():
+    |
+    return name
+```
+we construct an input to the model by
+
+1. Inserting `<mask_1>` token in place of cursor position
+2. Append `<sep>` token to indicate the boundary
+3. Insert another `<mask_1>` to indicate which mask we want to infill.
+
+The final snippet looks as follows:
+
+```python
+from transformers import AutoTokenizer, AutoModelForCausalLM
+tokenizer = AutoTokenizer.from_pretrained("Salesforce/codegen2-16B")
+model = AutoModelForCausalLM.from_pretrained("Salesforce/codegen2-16B", trust_remote_code=True, revision="main")
+
+
+def format(prefix, suffix):
+  return prefix + "<mask_1>" + suffix + "<|endoftext|>" + "<sep>" + "<mask_1>"
+
+
+prefix = "def hello_world():
+    "
+suffix = "    return name"
+text = format(prefix, suffix)
+input_ids = tokenizer(text, return_tensors="pt").input_ids
+generated_ids = model.generate(input_ids, max_length=128)
+print(tokenizer.decode(generated_ids[0], skip_special_tokens=False)[len(text):])
+```
+
+You might want to truncate the model output with `<eom>`.
+
+## Training data
+
+This checkpoint is trained on the stricter permissive subset of [the deduplicated version of the Stack dataset (v1.1)](https://huggingface.co/datasets/bigcode/the-stack-dedup). Supported languages (and frameworks) are as follows:
+`c`, `c++`, `c-sharp`, `dart`, `go`, `java`, `javascript`, `kotlin`, `lua`, `php`, `python`, `ruby`, `rust`, `scala`, `shell`, `sql`, `swift`, `typescript`, `vue`.
+
+## Training procedure
+
+CodeGen2 was trained using cross-entropy loss to maximize the likelihood of sequential inputs.
+The input sequences are formatted in two ways: (1) causal language modeling and (2) file-level span corruption.
+Please refer to the paper for more details.
+
+## Evaluation results
+
+We evaluate our models on HumanEval and HumanEval-Infill. Please refer to the [paper](https://arxiv.org/abs/2305.02309) for more details.
+
+## Intended use and limitations
+
+As an autoregressive language model, CodeGen2 is capable of extracting features from given natural language and programming language texts, and calculating the likelihood of them.
+However, the model is intended for and best at **program synthesis**, that is, generating executable code given English prompts, where the prompts should be in the form of a comment string. The model can complete partially-generated code as well.
+
+
+## BibTeX entry and citation info
+
+```bibtex
+@article{Nijkamp2023codegen2,
+  title={CodeGen2: Lessons for Training LLMs on Programming and Natural Languages},
+  author={Nijkamp, Erik and Hayashi, Hiroaki and Xiong, Caiming and Savarese, Silvio and Zhou, Yingbo},
+  journal={arXiv preprint},
+  year={2023}
+}
+```
+    

config.json

@@ -0,0 +1,27 @@
+{
+  "activation_function": "gelu_new",
+  "architectures": [
+    "GPTJForCausalLM"
+  ],
+  "attn_pdrop": 0.0,
+  "bos_token_id": 1,
+  "embd_pdrop": 0.0,
+  "eos_token_id": 2,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "gptj",
+  "n_embd": 4096,
+  "n_head": 16,
+  "n_inner": null,
+  "n_layer": 32,
+  "n_positions": 2048,
+  "resid_pdrop": 0.0,
+  "rotary_dim": 64,
+  "scale_attn_weights": true,
+  "tie_word_embeddings": false,
+  "tokenizer_class": "CodeGenTokenizer",
+  "torch_dtype": "float16",
+  "transformers_version": "4.28.1",
+  "use_cache": true,
+  "vocab_size": 51200
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "transformers_version": "4.28.1"
+}

pytorch_model-00001-of-00002.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef1c3fa0255c2516ed29df4b40a2656fb983d945e607400267ff954f0520aa44
+size 10051017272

pytorch_model-00002-of-00002.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b739f55967d28ae465d70a986875ba72f9e5e39b242d08cf9066e4ebd7d4ab15
+size 3809047229

pytorch_model.bin.index.json

@@ -0,0 +1,396 @@
+{
+  "metadata": {
+    "total_size": 13742493760.0
+  },
+  "weight_map": {
+    "lm_head.bias": "pytorch_model-00002-of-00002.bin",
+    "lm_head.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.0.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.0.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.0.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.0.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.0.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.0.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.0.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.0.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.0.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.0.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.0.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.0.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.1.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.1.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.1.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.1.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.1.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.1.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.1.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.1.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.1.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.1.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.1.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.1.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.10.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.10.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.10.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.10.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.10.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.10.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.10.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.10.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.10.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.10.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.10.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.10.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.11.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.11.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.11.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.11.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.11.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.11.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.11.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.11.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.11.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.11.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.11.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.11.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.12.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.12.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.12.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.12.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.12.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.12.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.12.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.12.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.12.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.12.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.12.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.12.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.13.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.13.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.13.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.13.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.13.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.13.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.13.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.13.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.13.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.13.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.13.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.13.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.14.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.14.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.14.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.14.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.14.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.14.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.14.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.14.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.14.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.14.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.14.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.14.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.15.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.15.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.15.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.15.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.15.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.15.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.15.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.15.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.15.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.15.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.15.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.15.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.16.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.16.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.16.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.16.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.16.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.16.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.16.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.16.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.16.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.16.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.16.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.16.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.17.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.17.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.17.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.17.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.17.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.17.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.17.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.17.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.17.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.17.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.17.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.17.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.18.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.18.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.18.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.18.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.18.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.18.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.18.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.18.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.18.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.18.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.18.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.18.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.19.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.19.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.19.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.19.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.19.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.19.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.19.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.19.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.19.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.19.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.19.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.19.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.2.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.2.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.2.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.2.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.2.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.2.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.2.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.2.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.2.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.2.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.2.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.2.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.20.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.20.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.20.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.20.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.20.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.20.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.20.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.20.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.20.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.20.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.20.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.20.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.21.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.21.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.21.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.21.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.21.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.21.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.21.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.21.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.21.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.21.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.21.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.21.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.22.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.22.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.22.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.22.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.22.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.22.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.22.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.22.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.22.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.22.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.22.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.22.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.23.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.23.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.23.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.23.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.23.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.23.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.23.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.23.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.23.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.23.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.23.mlp.fc_out.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.23.mlp.fc_out.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.24.attn.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.24.attn.k_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.24.attn.masked_bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.24.attn.out_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.24.attn.q_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.24.attn.v_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.24.ln_1.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.24.ln_1.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.24.mlp.fc_in.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.24.mlp.fc_in.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.24.mlp.fc_out.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.24.mlp.fc_out.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.25.attn.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.25.attn.k_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.25.attn.masked_bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.25.attn.out_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.25.attn.q_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.25.attn.v_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.25.ln_1.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.25.ln_1.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.25.mlp.fc_in.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.25.mlp.fc_in.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.25.mlp.fc_out.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.25.mlp.fc_out.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.26.attn.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.26.attn.k_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.26.attn.masked_bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.26.attn.out_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.26.attn.q_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.26.attn.v_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.26.ln_1.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.26.ln_1.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.26.mlp.fc_in.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.26.mlp.fc_in.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.26.mlp.fc_out.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.26.mlp.fc_out.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.27.attn.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.27.attn.k_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.27.attn.masked_bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.27.attn.out_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.27.attn.q_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.27.attn.v_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.27.ln_1.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.27.ln_1.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.27.mlp.fc_in.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.27.mlp.fc_in.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.27.mlp.fc_out.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.27.mlp.fc_out.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.28.attn.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.28.attn.k_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.28.attn.masked_bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.28.attn.out_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.28.attn.q_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.28.attn.v_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.28.ln_1.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.28.ln_1.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.28.mlp.fc_in.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.28.mlp.fc_in.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.28.mlp.fc_out.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.28.mlp.fc_out.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.29.attn.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.29.attn.k_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.29.attn.masked_bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.29.attn.out_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.29.attn.q_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.29.attn.v_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.29.ln_1.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.29.ln_1.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.29.mlp.fc_in.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.29.mlp.fc_in.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.29.mlp.fc_out.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.29.mlp.fc_out.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.3.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.3.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.3.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.3.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.3.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.3.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.3.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.3.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.3.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.3.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.3.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.3.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.30.attn.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.30.attn.k_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.30.attn.masked_bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.30.attn.out_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.30.attn.q_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.30.attn.v_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.30.ln_1.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.30.ln_1.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.30.mlp.fc_in.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.30.mlp.fc_in.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.30.mlp.fc_out.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.30.mlp.fc_out.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.31.attn.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.31.attn.k_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.31.attn.masked_bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.31.attn.out_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.31.attn.q_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.31.attn.v_proj.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.31.ln_1.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.31.ln_1.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.31.mlp.fc_in.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.31.mlp.fc_in.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.31.mlp.fc_out.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.31.mlp.fc_out.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.h.4.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.4.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.4.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.4.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.4.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.4.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.4.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.4.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.4.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.4.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.4.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.4.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.5.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.5.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.5.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.5.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.5.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.5.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.5.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.5.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.5.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.5.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.5.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.5.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.6.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.6.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.6.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.6.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.6.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.6.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.6.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.6.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.6.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.6.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.6.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.6.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.7.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.7.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.7.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.7.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.7.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.7.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.7.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.7.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.7.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.7.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.7.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.7.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.8.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.8.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.8.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.8.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.8.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.8.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.8.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.8.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.8.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.8.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.8.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.8.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.9.attn.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.9.attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.9.attn.masked_bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.9.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.9.attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.9.attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.9.ln_1.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.9.ln_1.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.9.mlp.fc_in.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.9.mlp.fc_in.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.9.mlp.fc_out.bias": "pytorch_model-00001-of-00002.bin",
+    "transformer.h.9.mlp.fc_out.weight": "pytorch_model-00001-of-00002.bin",
+    "transformer.ln_f.bias": "pytorch_model-00002-of-00002.bin",
+    "transformer.ln_f.weight": "pytorch_model-00002-of-00002.bin",
+    "transformer.wte.weight": "pytorch_model-00001-of-00002.bin"
+  }
+}