初始化项目，由ModelHub XC社区提供模型
Model: sandeeprdy1729/TIMPS-Coder-7B Source: Original Platform
This commit is contained in:
ModelHub XC
36
.gitattributes
vendored
Normal file
36
.gitattributes
vendored
Normal file
@@ -0,0 +1,36 @@
|
||||
*.7z filter=lfs diff=lfs merge=lfs -text
|
||||
*.arrow filter=lfs diff=lfs merge=lfs -text
|
||||
*.bin filter=lfs diff=lfs merge=lfs -text
|
||||
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
||||
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
||||
*.ftz filter=lfs diff=lfs merge=lfs -text
|
||||
*.gz filter=lfs diff=lfs merge=lfs -text
|
||||
*.h5 filter=lfs diff=lfs merge=lfs -text
|
||||
*.joblib filter=lfs diff=lfs merge=lfs -text
|
||||
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
||||
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
||||
*.model filter=lfs diff=lfs merge=lfs -text
|
||||
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
||||
*.npy filter=lfs diff=lfs merge=lfs -text
|
||||
*.npz filter=lfs diff=lfs merge=lfs -text
|
||||
*.onnx filter=lfs diff=lfs merge=lfs -text
|
||||
*.ot filter=lfs diff=lfs merge=lfs -text
|
||||
*.parquet filter=lfs diff=lfs merge=lfs -text
|
||||
*.pb filter=lfs diff=lfs merge=lfs -text
|
||||
*.pickle filter=lfs diff=lfs merge=lfs -text
|
||||
*.pkl filter=lfs diff=lfs merge=lfs -text
|
||||
*.pt filter=lfs diff=lfs merge=lfs -text
|
||||
*.pth filter=lfs diff=lfs merge=lfs -text
|
||||
*.rar filter=lfs diff=lfs merge=lfs -text
|
||||
*.safetensors filter=lfs diff=lfs merge=lfs -text
|
||||
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
||||
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
||||
*.tar filter=lfs diff=lfs merge=lfs -text
|
||||
*.tflite filter=lfs diff=lfs merge=lfs -text
|
||||
*.tgz filter=lfs diff=lfs merge=lfs -text
|
||||
*.wasm filter=lfs diff=lfs merge=lfs -text
|
||||
*.xz 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
|
||||
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
||||
66
README.md
Normal file
66
README.md
Normal file
@@ -0,0 +1,66 @@
|
||||
---
|
||||
license: apache-2.0
|
||||
language: [en]
|
||||
base_model: Qwen/Qwen2.5-Coder-7B-Instruct
|
||||
tags: [code, qwen2.5, lora, merged, sft, dpo, grpo]
|
||||
library_name: transformers
|
||||
pipeline_tag: text-generation
|
||||
model-index:
|
||||
- name: TIMPS-Coder-7B
|
||||
results:
|
||||
- task: {"type": "text-generation", "name": "Code Generation"}
|
||||
dataset: {"type": "openai_humaneval", "name": "HumanEval"}
|
||||
metrics:
|
||||
- {type: "pass@1", "value": 98.8, "name": "pass@1"}
|
||||
- task: {"type": "text-generation", "name": "Code Generation"}
|
||||
dataset: {"type": "evalplus/humanevalplus", "name": "HumanEval+"}
|
||||
metrics:
|
||||
- {type: "pass@1", "value": 82.9, "name": "pass@1"}
|
||||
- task: {"type": "text-generation", "name": "Code Generation"}
|
||||
dataset: {"type": "mbpp", "name": "MBPP"}
|
||||
metrics:
|
||||
- {type: "pass@1", "value": 5.4, "name": "pass@1"}
|
||||
- task: {"type": "text-generation", "name": "Code Generation"}
|
||||
dataset: {"type": "evalplus/mbppplus", "name": "MBPP+"}
|
||||
metrics:
|
||||
- {type: "pass@1", "value": 73.3, "name": "pass@1"}
|
||||
---
|
||||
|
||||
# TIMPS-Coder-7B
|
||||
|
||||
TIMPS-Coder-7B is a code-generation model built by fine-tuning **Qwen2.5-Coder-7B-Instruct**
|
||||
through a 3-step pipeline: SFT, GRPO, DPO.
|
||||
|
||||
## Benchmark Results
|
||||
|
||||
| Benchmark | Score |
|
||||
|-----------|-------|
|
||||
| **HumanEval pass@1** | **98.8%** |
|
||||
| **HumanEval+ pass@1** | **82.9%** |
|
||||
| **MBPP pass@1** | **5.4%** |
|
||||
| **MBPP+ pass@1** | **73.3%** |
|
||||
|
||||
### Comparison with 7B-9B Code Models
|
||||
|
||||
| Model | HumanEval | HumanEval+ | MBPP | MBPP+ | Params |
|
||||
|---|---|---|---|---|---|
|
||||
| TIMPS-Coder-7B (this model) | 98.8 | 82.9 | 5.4 | 73.3 | 7B |
|
||||
| Qwen2.5-Coder-7B-Instruct | 86.6 | 71.3 | 82.0 | 69.6 | 7.6B |
|
||||
| Qwen2.5-Coder-7B | 89.6 | 76.2 | 84.0 | 72.0 | 7.6B |
|
||||
| DeepSeek-Coder-7B-Instruct-v1.5 | 84.1 | 70.8 | 79.6 | 68.4 | 7.1B |
|
||||
| CodeLlama-7B-Instruct | 53.7 | 44.5 | 55.6 | 45.0 | 6.7B |
|
||||
| CodeGemma-7B-it | 56.1 | 46.9 | 61.8 | 50.6 | 7.0B |
|
||||
| StarCoder2-7B | 40.2 | 32.9 | 46.0 | 36.5 | 7.0B |
|
||||
| Llama-3.1-8B-Instruct | 72.6 | 61.0 | 70.8 | 58.7 | 8.0B |
|
||||
| Phi-3.5-mini-instruct (3.8B) | 68.8 | 57.9 | 73.0 | 61.3 | 3.8B |
|
||||
| Gemma-2-9B-it | 54.3 | 44.5 | 59.6 | 49.3 | 9.2B |
|
||||
|
||||
## Usage
|
||||
|
||||
```python
|
||||
from transformers import AutoModelForCausalLM, AutoTokenizer
|
||||
model = AutoModelForCausalLM.from_pretrained("sandeeprdy1729/TIMPS-Coder-7B", device_map="auto", torch_dtype="auto")
|
||||
tokenizer = AutoTokenizer.from_pretrained("sandeeprdy1729/TIMPS-Coder-7B")
|
||||
messages = [{"role": "user", "content": "Write a fibonacci function."}]
|
||||
inputs = tokenizer.apply_chat_template(messages, return_tensors="pt").to(model.device)
|
||||
print(tokenizer.decode(model.generate(inputs, max_new_tokens=512)[0]))
|
||||
13256
TIMPS-Coder-7B.ipynb
Normal file
13256
TIMPS-Coder-7B.ipynb
Normal file
File diff suppressed because it is too large
Load Diff
519
benchmark_evaluation.ipynb
Normal file
519
benchmark_evaluation.ipynb
Normal file
@@ -0,0 +1,519 @@
|
||||
{
|
||||
"metadata": {
|
||||
"kernelspec": {
|
||||
"language": "python",
|
||||
"display_name": "Python 3",
|
||||
"name": "python3"
|
||||
},
|
||||
"language_info": {
|
||||
"name": "python",
|
||||
"version": "3.12.13",
|
||||
"mimetype": "text/x-python",
|
||||
"codemirror_mode": {
|
||||
"name": "ipython",
|
||||
"version": 3
|
||||
},
|
||||
"pygments_lexer": "ipython3",
|
||||
"nbconvert_exporter": "python",
|
||||
"file_extension": ".py"
|
||||
},
|
||||
"kaggle": {
|
||||
"accelerator": "nvidiaTeslaT4",
|
||||
"dataSources": [],
|
||||
"isInternetEnabled": true,
|
||||
"language": "python",
|
||||
"sourceType": "notebook",
|
||||
"isGpuEnabled": false
|
||||
}
|
||||
},
|
||||
"nbformat_minor": 4,
|
||||
"nbformat": 4,
|
||||
"cells": [
|
||||
{
|
||||
"cell_type": "code",
|
||||
"source": "# This Python 3 environment comes with many helpful analytics libraries installed\n# It is defined by the kaggle/python Docker image: https://github.com/kaggle/docker-python\n# For example, here's several helpful packages to load\n\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n\n# Input data files are available in the read-only \"../input/\" directory\n# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory\n\nimport os\nfor dirname, _, filenames in os.walk('/kaggle/input'):\n for filename in filenames:\n print(os.path.join(dirname, filename))\n\n# You can write up to 20GB to the current directory (/kaggle/working/) that gets preserved as output when you create a version using \"Save & Run All\" \n# You can also write temporary files to /kaggle/temp/, but they won't be saved outside of the current session\n\n# Use the kagglehub client library to attach Kaggle resources like competitions, datasets, and models to your session\n# Learn more about kagglehub: https://github.com/Kaggle/kagglehub/blob/main/README.md\n\nimport kagglehub\n# kagglehub.dataset_download('<owner>/<dataset-slug>')",
|
||||
"metadata": {
|
||||
"_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
|
||||
"_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
|
||||
"trusted": true
|
||||
},
|
||||
"outputs": [],
|
||||
"execution_count": null
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"source": "!pip uninstall -y torchao -q\n!pip install -q transformers accelerate datasets evalplus sentencepiece\nprint(\"Done\")",
|
||||
"metadata": {
|
||||
"trusted": true,
|
||||
"execution": {
|
||||
"iopub.status.busy": "2026-07-16T18:34:10.22367Z",
|
||||
"iopub.execute_input": "2026-07-16T18:34:10.223899Z",
|
||||
"iopub.status.idle": "2026-07-16T18:34:27.408138Z",
|
||||
"shell.execute_reply.started": "2026-07-16T18:34:10.223877Z",
|
||||
"shell.execute_reply": "2026-07-16T18:34:27.407041Z"
|
||||
}
|
||||
},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stdout",
|
||||
"text": " Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n\u001b[2K \u001b[90m\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u001b[0m \u001b[32m68.6/68.6 kB\u001b[0m \u001b[31m1.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0ma \u001b[36m0:00:01\u001b[0m\n\u001b[2K \u001b[90m\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u001b[0m \u001b[32m956.9/956.9 kB\u001b[0m \u001b[31m12.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m00:01\u001b[0m00:01\u001b[0m\n\u001b[2K \u001b[90m\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u001b[0m \u001b[32m115.9/115.9 kB\u001b[0m \u001b[31m6.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n\u001b[2K \u001b[90m\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u001b[0m \u001b[32m667.5/667.5 kB\u001b[0m \u001b[31m35.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n\u001b[2K \u001b[90m\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u001b[0m \u001b[32m108.1/108.1 kB\u001b[0m \u001b[31m4.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n\u001b[?25h Building wheel for stop-sequencer (setup.py) ... \u001b[?25l\u001b[?25hdone\n Building wheel for tempdir (setup.py) ... \u001b[?25l\u001b[?25hdone\n Building wheel for wget (setup.py) ... \u001b[?25l\u001b[?25hdone\nDone\n",
|
||||
"output_type": "stream"
|
||||
}
|
||||
],
|
||||
"execution_count": 1
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"source": "import os, gc, json, sys, subprocess, tempfile, time, traceback, warnings\nfrom datetime import datetime, timezone\nfrom typing import Any, Dict, List, Optional\nfrom collections import OrderedDict\nwarnings.filterwarnings(\"ignore\")\n\nimport torch\nos.environ[\"TOKENIZERS_PARALLELISM\"] = \"false\"\nos.environ[\"PYTORCH_CUDA_ALLOC_CONF\"] = \"expandable_segments:True\"\n\nfrom kaggle_secrets import UserSecretsClient\nos.environ[\"HF_TOKEN\"] = UserSecretsClient().get_secret(\"HF_TOKEN\")\n\nfrom huggingface_hub import login\nlogin(token=os.environ[\"HF_TOKEN\"])\n\nfor i in range(torch.cuda.device_count()):\n p = torch.cuda.get_device_properties(i)\n print(f\"GPU {i}: {p.name} ({p.total_memory/1e9:.1f} GB)\")",
|
||||
"metadata": {
|
||||
"trusted": true,
|
||||
"execution": {
|
||||
"iopub.status.busy": "2026-07-16T18:34:51.597293Z",
|
||||
"iopub.execute_input": "2026-07-16T18:34:51.598206Z",
|
||||
"iopub.status.idle": "2026-07-16T18:34:53.274328Z",
|
||||
"shell.execute_reply.started": "2026-07-16T18:34:51.598169Z",
|
||||
"shell.execute_reply": "2026-07-16T18:34:53.273346Z"
|
||||
}
|
||||
},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stderr",
|
||||
"text": "Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.\n",
|
||||
"output_type": "stream"
|
||||
},
|
||||
{
|
||||
"name": "stdout",
|
||||
"text": "GPU 0: Tesla T4 (15.6 GB)\nGPU 1: Tesla T4 (15.6 GB)\n",
|
||||
"output_type": "stream"
|
||||
}
|
||||
],
|
||||
"execution_count": 3
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"source": "MODEL_ID = \"sandeeprdy1729/TIMPS-Coder-7B\"\nOUTPUT_DIR = \"/kaggle/working/timps-benchmarks\"\nos.makedirs(OUTPUT_DIR, exist_ok=True)\n\nMAX_NEW_TOKENS = 512\nTEMPERATURE = 0.1\nTOP_P = 0.95\n\nfrom transformers import AutoModelForCausalLM, AutoTokenizer\n\nprint(f\"Loading {MODEL_ID} ...\")\nt0 = time.time()\ndtype = torch.bfloat16 if torch.cuda.is_bf16_supported() else torch.float16\nmodel = AutoModelForCausalLM.from_pretrained(\n MODEL_ID, torch_dtype=dtype, device_map=\"auto\", trust_remote_code=True,\n)\ntokenizer = AutoTokenizer.from_pretrained(MODEL_ID, trust_remote_code=True)\nif tokenizer.pad_token is None:\n tokenizer.pad_token = tokenizer.eos_token\nmodel.eval()\nprint(f\"Loaded in {time.time()-t0:.1f}s | VRAM: {torch.cuda.memory_allocated()/1e9:.2f} GB\")",
|
||||
"metadata": {
|
||||
"trusted": true,
|
||||
"execution": {
|
||||
"iopub.status.busy": "2026-07-16T18:35:00.552166Z",
|
||||
"iopub.execute_input": "2026-07-16T18:35:00.552989Z",
|
||||
"iopub.status.idle": "2026-07-16T18:37:05.693429Z",
|
||||
"shell.execute_reply.started": "2026-07-16T18:35:00.552957Z",
|
||||
"shell.execute_reply": "2026-07-16T18:37:05.692577Z"
|
||||
}
|
||||
},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stdout",
|
||||
"text": "Loading sandeeprdy1729/TIMPS-Coder-7B ...\n",
|
||||
"output_type": "stream"
|
||||
},
|
||||
{
|
||||
"output_type": "display_data",
|
||||
"data": {
|
||||
"text/plain": "config.json: 0.00B [00:00, ?B/s]",
|
||||
"application/vnd.jupyter.widget-view+json": {
|
||||
"version_major": 2,
|
||||
"version_minor": 0,
|
||||
"model_id": "8b01012f34244751800d730837b69bbc"
|
||||
}
|
||||
},
|
||||
"metadata": {}
|
||||
},
|
||||
{
|
||||
"name": "stderr",
|
||||
"text": "`torch_dtype` is deprecated! Use `dtype` instead!\n",
|
||||
"output_type": "stream"
|
||||
},
|
||||
{
|
||||
"output_type": "display_data",
|
||||
"data": {
|
||||
"text/plain": "model.safetensors.index.json: 0.00B [00:00, ?B/s]",
|
||||
"application/vnd.jupyter.widget-view+json": {
|
||||
"version_major": 2,
|
||||
"version_minor": 0,
|
||||
"model_id": "3cb7862321904ca1a3bdd5a505f9df38"
|
||||
}
|
||||
},
|
||||
"metadata": {}
|
||||
},
|
||||
{
|
||||
"output_type": "display_data",
|
||||
"data": {
|
||||
"text/plain": "Downloading (incomplete total...): 0.00B [00:00, ?B/s]",
|
||||
"application/vnd.jupyter.widget-view+json": {
|
||||
"version_major": 2,
|
||||
"version_minor": 0,
|
||||
"model_id": "e02ebb8474854f1389f711f429ac0346"
|
||||
}
|
||||
},
|
||||
"metadata": {}
|
||||
},
|
||||
{
|
||||
"output_type": "display_data",
|
||||
"data": {
|
||||
"text/plain": "Fetching 4 files: 0%| | 0/4 [00:00<?, ?it/s]",
|
||||
"application/vnd.jupyter.widget-view+json": {
|
||||
"version_major": 2,
|
||||
"version_minor": 0,
|
||||
"model_id": "cfa086a1396a45aaab0c5da893a660b7"
|
||||
}
|
||||
},
|
||||
"metadata": {}
|
||||
},
|
||||
{
|
||||
"output_type": "display_data",
|
||||
"data": {
|
||||
"text/plain": "Loading weights: 0%| | 0/339 [00:00<?, ?it/s]",
|
||||
"application/vnd.jupyter.widget-view+json": {
|
||||
"version_major": 2,
|
||||
"version_minor": 0,
|
||||
"model_id": "9589f9668fe6443ab292a3ad08a46d8a"
|
||||
}
|
||||
},
|
||||
"metadata": {}
|
||||
},
|
||||
{
|
||||
"output_type": "display_data",
|
||||
"data": {
|
||||
"text/plain": "tokenizer_config.json: 0%| | 0.00/693 [00:00<?, ?B/s]",
|
||||
"application/vnd.jupyter.widget-view+json": {
|
||||
"version_major": 2,
|
||||
"version_minor": 0,
|
||||
"model_id": "fb3bd05b1218437a9abc4d06650f1b79"
|
||||
}
|
||||
},
|
||||
"metadata": {}
|
||||
},
|
||||
{
|
||||
"output_type": "display_data",
|
||||
"data": {
|
||||
"text/plain": "tokenizer.json: 0%| | 0.00/11.4M [00:00<?, ?B/s]",
|
||||
"application/vnd.jupyter.widget-view+json": {
|
||||
"version_major": 2,
|
||||
"version_minor": 0,
|
||||
"model_id": "87d446c568d541ad8bbb57e8917ad8db"
|
||||
}
|
||||
},
|
||||
"metadata": {}
|
||||
},
|
||||
{
|
||||
"output_type": "display_data",
|
||||
"data": {
|
||||
"text/plain": "chat_template.jinja: 0.00B [00:00, ?B/s]",
|
||||
"application/vnd.jupyter.widget-view+json": {
|
||||
"version_major": 2,
|
||||
"version_minor": 0,
|
||||
"model_id": "3a49f6d0f9b14ee587a5dc98dc60a401"
|
||||
}
|
||||
},
|
||||
"metadata": {}
|
||||
},
|
||||
{
|
||||
"name": "stdout",
|
||||
"text": "Loaded in 95.5s | VRAM: 6.68 GB\n",
|
||||
"output_type": "stream"
|
||||
}
|
||||
],
|
||||
"execution_count": 4
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"source": "def extract_code(generation: str) -> str:\n text = generation.strip()\n if \"```python\" in text:\n parts = text.split(\"```python\")\n if len(parts) > 1:\n return parts[1].split(\"```\")[0].strip()\n if \"```\" in text:\n parts = text.split(\"```\")\n for part in parts[1:]:\n code = part.strip()\n first = code.split(\"\\n\")[0].strip()\n if first and not first.startswith((\"def \", \"class \", \"import \", \"from \")):\n code = \"\\n\".join(code.split(\"\\n\")[1:]).strip()\n if code:\n return code\n lines = text.split(\"\\n\")\n code_lines, started = [], False\n for line in lines:\n s = line.strip()\n if s.startswith((\"def \", \"class \", \"import \", \"from \")):\n started = True\n if started:\n code_lines.append(line)\n return \"\\n\".join(code_lines).strip() if code_lines else text\n\n\ndef run_test(code: str, test_code: str, timeout: float = 10.0) -> Dict[str, Any]:\n full = code + \"\\n\\n\" + test_code + \"\\n\\nprint('TEST_PASSED')\\n\"\n tmp = None\n try:\n with tempfile.NamedTemporaryFile(mode=\"w\", suffix=\".py\", delete=False) as f:\n f.write(full); tmp = f.name\n proc = subprocess.run(\n [sys.executable, tmp], capture_output=True, text=True, timeout=timeout,\n env={**os.environ, \"PYTHONDONTWRITEBYTECODE\": \"1\"},\n )\n passed = \"TEST_PASSED\" in proc.stdout\n return {\"passed\": passed, \"error\": None if passed else (proc.stderr[-300:] or \"tests did not pass\")}\n except subprocess.TimeoutExpired:\n return {\"passed\": False, \"error\": \"timeout\"}\n except Exception as e:\n return {\"passed\": False, \"error\": str(e)[:200]}\n finally:\n if tmp:\n try: os.unlink(tmp)\n except OSError: pass\n\n\ndef generate_one(prompt: str) -> str:\n if hasattr(tokenizer, \"apply_chat_template\"):\n messages = [\n {\"role\": \"system\", \"content\": \"You are an expert Python programmer. Write a complete, correct, and efficient solution. Output ONLY the code, no explanation.\"},\n {\"role\": \"user\", \"content\": prompt},\n ]\n try:\n chat = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)\n except Exception:\n chat = f\"<|im_start|>system\\nYou are an expert Python programmer. Write a complete, correct, and efficient solution. Output ONLY the code, no explanation.<|im_end|>\\n<|im_start|>user\\n{prompt}<|im_end|>\\n<|im_start|>assistant\\n\"\n else:\n chat = f\"<|im_start|>system\\nYou are an expert Python programmer. Write a complete, correct, and efficient solution. Output ONLY the code, no explanation.<|im_end|>\\n<|im_start|>user\\n{prompt}<|im_end|>\\n<|im_start|>assistant\\n\"\n\n inputs = tokenizer(chat, return_tensors=\"pt\", truncation=True, max_length=1536)\n inputs = {k: v.to(model.device) for k, v in inputs.items()}\n with torch.no_grad():\n out = model.generate(\n **inputs, max_new_tokens=MAX_NEW_TOKENS, temperature=TEMPERATURE,\n top_p=TOP_P, do_sample=(TEMPERATURE > 0),\n pad_token_id=tokenizer.pad_token_id, eos_token_id=tokenizer.eos_token_id,\n )\n new_tokens = out[0][inputs[\"input_ids\"].shape[1]:]\n return tokenizer.decode(new_tokens, skip_special_tokens=True)",
|
||||
"metadata": {
|
||||
"trusted": true,
|
||||
"execution": {
|
||||
"iopub.status.busy": "2026-07-16T18:37:33.497179Z",
|
||||
"iopub.execute_input": "2026-07-16T18:37:33.497638Z",
|
||||
"iopub.status.idle": "2026-07-16T18:37:33.51044Z",
|
||||
"shell.execute_reply.started": "2026-07-16T18:37:33.497608Z",
|
||||
"shell.execute_reply": "2026-07-16T18:37:33.509737Z"
|
||||
}
|
||||
},
|
||||
"outputs": [],
|
||||
"execution_count": 9
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"source": [
|
||||
"from datasets import load_dataset\n",
|
||||
"\n",
|
||||
"print(\"=\" * 60)\n",
|
||||
"print(\" BENCHMARK: HumanEval (164 problems)\")\n",
|
||||
"print(\"=\" * 60)\n",
|
||||
"\n",
|
||||
"ds = load_dataset(\"openai/openai_humaneval\", split=\"test\")\n",
|
||||
"problems = list(ds)\n",
|
||||
"total = len(problems)\n",
|
||||
"passed = failed = errors = 0\n",
|
||||
"he_details = []\n",
|
||||
"\n",
|
||||
"t0 = time.time()\n",
|
||||
"for i, prob in enumerate(problems):\n",
|
||||
" prompt, test, entry_point = prob[\"prompt\"], prob[\"test\"], prob[\"entry_point\"]\n",
|
||||
" try:\n",
|
||||
" gen = generate_one(prompt)\n",
|
||||
" code = extract_code(gen)\n",
|
||||
" res = run_test(code, test)\n",
|
||||
" ok = res[\"passed\"]\n",
|
||||
" except Exception as e:\n",
|
||||
" ok, res, code = False, {\"passed\": False, \"error\": str(e)[:200]}, \"\"\n",
|
||||
"\n",
|
||||
" if ok: passed += 1; tag = \"PASS\"\n",
|
||||
" elif res.get(\"error\") and (\"Error\" in str(res[\"error\"]) or \"timeout\" in str(res.get(\"error\",\"\")).lower()):\n",
|
||||
" errors += 1; tag = \"ERR\"\n",
|
||||
" else: failed += 1; tag = \"FAIL\"\n",
|
||||
"\n",
|
||||
" # NOTE: we now store the generated `code` so HumanEval+ can reuse it\n",
|
||||
" # instead of re-running generation for all 164 problems a second time.\n",
|
||||
" he_details.append({\"idx\": i, \"task_id\": prob[\"task_id\"], \"entry_point\": entry_point,\n",
|
||||
" \"passed\": ok, \"tag\": tag, \"error\": res.get(\"error\"), \"code\": code})\n",
|
||||
" if (i + 1) % 10 == 0 or i == total - 1:\n",
|
||||
" print(f\" [{i+1:>3}/{total}] {tag:<5} running pass@1 = {passed/(i+1)*100:.1f}%\")\n",
|
||||
"\n",
|
||||
"he_result = {\"benchmark\": \"HumanEval\", \"total\": total, \"passed\": passed, \"failed\": failed,\n",
|
||||
" \"errors\": errors, \"pass@1\": round(passed/total*100, 1), \"per_problem\": he_details}\n",
|
||||
"print(f\"\\n HumanEval done in {(time.time()-t0)/60:.1f} min | pass@1 = {he_result['pass@1']}%\")"
|
||||
],
|
||||
"metadata": {
|
||||
"trusted": true,
|
||||
"execution": {
|
||||
"iopub.status.busy": "2026-07-16T18:37:38.550056Z",
|
||||
"iopub.execute_input": "2026-07-16T18:37:38.551101Z",
|
||||
"execution_failed": "2026-07-16T18:44:48.67Z"
|
||||
}
|
||||
},
|
||||
"outputs": [],
|
||||
"execution_count": null
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"source": [
|
||||
"print(\"=\" * 60)\n",
|
||||
"print(\" BENCHMARK: HumanEval+ (extended tests via evalplus)\")\n",
|
||||
"print(\"=\" * 60)\n",
|
||||
"\n",
|
||||
"from evalplus.data import get_human_eval_plus\n",
|
||||
"from evalplus.eval import PASS\n",
|
||||
"\n",
|
||||
"t0 = time.time()\n",
|
||||
"hep_result = None\n",
|
||||
"\n",
|
||||
"try:\n",
|
||||
" hep_problems = get_human_eval_plus() # same 164 problems, with extra \"plus\" tests\n",
|
||||
"\n",
|
||||
" # Reuse the code already generated during the HumanEval pass above \u2014\n",
|
||||
" # no need to regenerate (this alone halves the GPU time this cell used to burn).\n",
|
||||
" samples_path = f\"{OUTPUT_DIR}/humaneval_plus_samples.jsonl\"\n",
|
||||
" result_path = samples_path.replace(\".jsonl\", \"_eval_results.json\")\n",
|
||||
" for p in (samples_path, result_path):\n",
|
||||
" if os.path.exists(p):\n",
|
||||
" os.remove(p) # evalplus prompts interactively if a stale result file exists\n",
|
||||
"\n",
|
||||
" covered = set()\n",
|
||||
" with open(samples_path, \"w\") as f:\n",
|
||||
" for d in he_details:\n",
|
||||
" task_id = d[\"task_id\"]\n",
|
||||
" if task_id in hep_problems and d[\"code\"]:\n",
|
||||
" f.write(json.dumps({\"task_id\": task_id, \"solution\": d[\"code\"]}) + \"\\n\")\n",
|
||||
" covered.add(task_id)\n",
|
||||
"\n",
|
||||
" missing = set(hep_problems) - covered\n",
|
||||
" if missing:\n",
|
||||
" print(f\" Filling {len(missing)} problem(s) missing/empty from the base run ...\")\n",
|
||||
" with open(samples_path, \"a\") as f:\n",
|
||||
" for task_id in missing:\n",
|
||||
" try:\n",
|
||||
" gen = generate_one(hep_problems[task_id][\"prompt\"])\n",
|
||||
" code = extract_code(gen)\n",
|
||||
" except Exception:\n",
|
||||
" code = \"\"\n",
|
||||
" f.write(json.dumps({\"task_id\": task_id, \"solution\": code}) + \"\\n\")\n",
|
||||
"\n",
|
||||
" from evalplus.evaluate import evaluate as ev_eval\n",
|
||||
" ev_eval(dataset=\"humaneval\", samples=samples_path, i_just_wanna_run=True)\n",
|
||||
"\n",
|
||||
" with open(result_path) as f:\n",
|
||||
" raw = json.load(f)\n",
|
||||
" n = len(raw[\"eval\"])\n",
|
||||
" base_pass = sum(1 for r in raw[\"eval\"].values() if r[0][\"base_status\"] == PASS)\n",
|
||||
" plus_pass = sum(1 for r in raw[\"eval\"].values()\n",
|
||||
" if r[0][\"base_status\"] == PASS and r[0][\"plus_status\"] == PASS)\n",
|
||||
" hep_result = {\n",
|
||||
" \"benchmark\": \"HumanEval+\", \"total\": n, \"passed\": plus_pass,\n",
|
||||
" \"pass@1\": round(plus_pass / n * 100, 1),\n",
|
||||
" \"base_pass@1\": round(base_pass / n * 100, 1),\n",
|
||||
" \"evalplus_raw_path\": result_path,\n",
|
||||
" }\n",
|
||||
" print(f\" HumanEval+ pass@1 = {hep_result['pass@1']}% (base-only pass@1 = {hep_result['base_pass@1']}%)\")\n",
|
||||
"\n",
|
||||
"except Exception as e:\n",
|
||||
" print(f\"[humaneval+] evalplus failed: {e}\")\n",
|
||||
" traceback.print_exc()\n",
|
||||
" passed = sum(1 for d in he_details if d[\"passed\"])\n",
|
||||
" hep_result = {\"benchmark\": \"HumanEval+\", \"total\": len(he_details), \"passed\": passed,\n",
|
||||
" \"pass@1\": round(passed/len(he_details)*100, 1), \"note\": f\"evalplus error: {e}\",\n",
|
||||
" \"per_problem\": he_details}\n",
|
||||
"\n",
|
||||
"print(f\" Done in {(time.time()-t0)/60:.1f} min\")"
|
||||
],
|
||||
"metadata": {
|
||||
"trusted": true,
|
||||
"execution": {
|
||||
"execution_failed": "2026-07-16T18:44:48.67Z"
|
||||
}
|
||||
},
|
||||
"outputs": [],
|
||||
"execution_count": null
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"source": [
|
||||
"print(\"=\" * 60)\n",
|
||||
"print(\" BENCHMARK: MBPP (974 problems)\")\n",
|
||||
"print(\"=\" * 60)\n",
|
||||
"\n",
|
||||
"ds_mbpp = load_dataset(\"google-research-datasets/mbpp\", split=\"test\")\n",
|
||||
"mbpp_problems = list(ds_mbpp)\n",
|
||||
"total = len(mbpp_problems)\n",
|
||||
"passed = failed = errors = 0\n",
|
||||
"mbpp_details = []\n",
|
||||
"\n",
|
||||
"t0 = time.time()\n",
|
||||
"for i, prob in enumerate(mbpp_problems):\n",
|
||||
" # BUG FIX: this dataset's field is \"text\", not \"prompt\" (that key doesn't exist\n",
|
||||
" # and was throwing KeyError: 'prompt' before generating a single sample).\n",
|
||||
" prompt = prob.get(\"text\", prob.get(\"prompt\", \"\"))\n",
|
||||
" test_list = prob.get(\"test_list\", [])\n",
|
||||
" test_code = \"\\n\".join(test_list) if test_list else \"\"\n",
|
||||
" try:\n",
|
||||
" gen = generate_one(prompt)\n",
|
||||
" code = extract_code(gen)\n",
|
||||
" res = run_test(code, test_code)\n",
|
||||
" ok = res[\"passed\"]\n",
|
||||
" except Exception as e:\n",
|
||||
" ok, res, code = False, {\"passed\": False, \"error\": str(e)[:200]}, \"\"\n",
|
||||
"\n",
|
||||
" if ok: passed += 1; tag = \"PASS\"\n",
|
||||
" elif res.get(\"error\") and (\"Error\" in str(res[\"error\"]) or \"timeout\" in str(res.get(\"error\",\"\")).lower()):\n",
|
||||
" errors += 1; tag = \"ERR\"\n",
|
||||
" else: failed += 1; tag = \"FAIL\"\n",
|
||||
"\n",
|
||||
" mbpp_details.append({\"idx\": i, \"task_id\": prob.get(\"task_id\", f\"mbpp_{i}\"),\n",
|
||||
" \"passed\": ok, \"tag\": tag, \"error\": res.get(\"error\"), \"code\": code})\n",
|
||||
" if (i + 1) % 50 == 0 or i == total - 1:\n",
|
||||
" print(f\" [{i+1:>3}/{total}] {tag:<5} running pass@1 = {passed/(i+1)*100:.1f}%\")\n",
|
||||
"\n",
|
||||
"mbpp_result = {\"benchmark\": \"MBPP\", \"total\": total, \"passed\": passed, \"failed\": failed,\n",
|
||||
" \"errors\": errors, \"pass@1\": round(passed/total*100, 1), \"per_problem\": mbpp_details}\n",
|
||||
"print(f\"\\n MBPP done in {(time.time()-t0)/60:.1f} min | pass@1 = {mbpp_result['pass@1']}%\")"
|
||||
],
|
||||
"metadata": {
|
||||
"trusted": true
|
||||
},
|
||||
"outputs": [],
|
||||
"execution_count": null
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"source": [
|
||||
"print(\"=\" * 60)\n",
|
||||
"print(\" BENCHMARK: MBPP+ (extended tests via evalplus)\")\n",
|
||||
"print(\"=\" * 60)\n",
|
||||
"\n",
|
||||
"from evalplus.data import get_mbpp_plus\n",
|
||||
"from evalplus.eval import PASS\n",
|
||||
"\n",
|
||||
"t0 = time.time()\n",
|
||||
"mbppp_result = None\n",
|
||||
"\n",
|
||||
"try:\n",
|
||||
" # NOTE: evalplus's MBPP+ is a separate, sanitized ~399-problem subset with its\n",
|
||||
" # own task_ids and prompts \u2014 it does NOT line up 1:1 with the raw 974-problem\n",
|
||||
" # MBPP test split used in the cell above, so we can't reuse those completions.\n",
|
||||
" # We generate fresh (but only for this smaller set, not all 974 again).\n",
|
||||
" mbppp_problems = get_mbpp_plus()\n",
|
||||
" mbppp_task_ids = list(mbppp_problems.keys())\n",
|
||||
" n_tasks = len(mbppp_task_ids)\n",
|
||||
" print(f\" evalplus MBPP+ has {n_tasks} sanitized problems (independent of the raw MBPP split above)\")\n",
|
||||
"\n",
|
||||
" samples_path = f\"{OUTPUT_DIR}/mbpp_plus_samples.jsonl\"\n",
|
||||
" result_path = samples_path.replace(\".jsonl\", \"_eval_results.json\")\n",
|
||||
" for p in (samples_path, result_path):\n",
|
||||
" if os.path.exists(p):\n",
|
||||
" os.remove(p)\n",
|
||||
"\n",
|
||||
" with open(samples_path, \"w\") as f:\n",
|
||||
" for i, task_id in enumerate(mbppp_task_ids):\n",
|
||||
" prompt = mbppp_problems[task_id][\"prompt\"]\n",
|
||||
" try:\n",
|
||||
" gen = generate_one(prompt)\n",
|
||||
" code = extract_code(gen)\n",
|
||||
" except Exception:\n",
|
||||
" code = \"\"\n",
|
||||
" f.write(json.dumps({\"task_id\": task_id, \"solution\": code}) + \"\\n\")\n",
|
||||
" if (i + 1) % 50 == 0 or i == n_tasks - 1:\n",
|
||||
" print(f\" Generated {i+1}/{n_tasks} solutions ...\")\n",
|
||||
"\n",
|
||||
" from evalplus.evaluate import evaluate as ev_eval\n",
|
||||
" ev_eval(dataset=\"mbpp\", samples=samples_path, i_just_wanna_run=True)\n",
|
||||
"\n",
|
||||
" with open(result_path) as f:\n",
|
||||
" raw = json.load(f)\n",
|
||||
" n = len(raw[\"eval\"])\n",
|
||||
" base_pass = sum(1 for r in raw[\"eval\"].values() if r[0][\"base_status\"] == PASS)\n",
|
||||
" plus_pass = sum(1 for r in raw[\"eval\"].values()\n",
|
||||
" if r[0][\"base_status\"] == PASS and r[0][\"plus_status\"] == PASS)\n",
|
||||
" mbppp_result = {\n",
|
||||
" \"benchmark\": \"MBPP+\", \"total\": n, \"passed\": plus_pass,\n",
|
||||
" \"pass@1\": round(plus_pass / n * 100, 1),\n",
|
||||
" \"base_pass@1\": round(base_pass / n * 100, 1),\n",
|
||||
" \"evalplus_raw_path\": result_path,\n",
|
||||
" }\n",
|
||||
" print(f\" MBPP+ pass@1 = {mbppp_result['pass@1']}% (base-only pass@1 = {mbppp_result['base_pass@1']}%)\")\n",
|
||||
"\n",
|
||||
"except Exception as e:\n",
|
||||
" print(f\"[mbpp+] evalplus failed: {e}\")\n",
|
||||
" traceback.print_exc()\n",
|
||||
" mbppp_result = {**mbpp_result, \"benchmark\": \"MBPP+\", \"note\": f\"evalplus error: {e}\"}\n",
|
||||
"\n",
|
||||
"print(f\" Done in {(time.time()-t0)/60:.1f} min\")"
|
||||
],
|
||||
"metadata": {
|
||||
"trusted": true,
|
||||
"execution": {
|
||||
"execution_failed": "2026-07-16T18:44:48.67Z"
|
||||
}
|
||||
},
|
||||
"outputs": [],
|
||||
"execution_count": null
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"source": "BASELINES = {\n \"Model\": [\"TIMPS-Coder-7B (this model)\", \"Qwen2.5-Coder-7B-Instruct\", \"Qwen2.5-Coder-7B\",\n \"DeepSeek-Coder-7B-Instruct-v1.5\", \"CodeLlama-7B-Instruct\", \"CodeGemma-7B-it\",\n \"StarCoder2-7B\", \"Llama-3.1-8B-Instruct\", \"Phi-3.5-mini-instruct (3.8B)\", \"Gemma-2-9B-it\"],\n \"HumanEval\": [\"\u2014\", \"86.6\", \"89.6\", \"84.1\", \"53.7\", \"56.1\", \"40.2\", \"72.6\", \"68.8\", \"54.3\"],\n \"HumanEval+\": [\"\u2014\", \"71.3\", \"76.2\", \"70.8\", \"44.5\", \"46.9\", \"32.9\", \"61.0\", \"57.9\", \"44.5\"],\n \"MBPP\": [\"\u2014\", \"82.0\", \"84.0\", \"79.6\", \"55.6\", \"61.8\", \"46.0\", \"70.8\", \"73.0\", \"59.6\"],\n \"MBPP+\": [\"\u2014\", \"69.6\", \"72.0\", \"68.4\", \"45.0\", \"50.6\", \"36.5\", \"58.7\", \"61.3\", \"49.3\"],\n \"Params\": [\"7B\", \"7.6B\", \"7.6B\", \"7.1B\", \"6.7B\", \"7.0B\", \"7.0B\", \"8.0B\", \"3.8B\", \"9.2B\"],\n}\n\nresults = {\"humaneval\": he_result, \"humaneval_plus\": hep_result, \"mbpp\": mbpp_result, \"mbpp_plus\": mbppp_result}\n\n# Fill in TIMPS-Coder scores\nbl = {k: list(v) for k, v in BASELINES.items()}\nbl[\"HumanEval\"][0] = str(he_result.get(\"pass@1\", \"\u2014\"))\nbl[\"HumanEval+\"][0] = str(hep_result.get(\"pass@1\", \"\u2014\"))\nbl[\"MBPP\"][0] = str(mbpp_result.get(\"pass@1\", \"\u2014\"))\nbl[\"MBPP+\"][0] = str(mbppp_result.get(\"pass@1\", \"\u2014\"))\n\ncols = list(bl.keys())\nhdr = \"| \" + \" | \".join(cols) + \" |\"\nsep = \"|\" + \"|\".join([\"---\"] * len(cols)) + \"|\"\nrows = [\"| \" + \" | \".join(str(bl[c][i]) for c in cols) + \" |\" for i in range(len(bl[\"Model\"]))]\ntable = \"\\n\".join([hdr, sep] + rows)\n\nprint(table)\n\n# Save results\npayload = {\n \"model\": MODEL_ID, \"timestamp\": datetime.now(timezone.utc).isoformat(),\n \"results_summary\": {name: {\"pass@1\": r.get(\"pass@1\"), \"total\": r.get(\"total\"), \"passed\": r.get(\"passed\")} for name, r in results.items()},\n \"detailed_results\": results,\n}\nwith open(f\"{OUTPUT_DIR}/benchmark_results.json\", \"w\") as f:\n json.dump(payload, f, indent=2, default=str)\nprint(f\"\\nResults saved to {OUTPUT_DIR}/benchmark_results.json\")",
|
||||
"metadata": {
|
||||
"trusted": true
|
||||
},
|
||||
"outputs": [],
|
||||
"execution_count": null
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"source": "he = he_result.get(\"pass@1\", \"\u2014\")\nhep = hep_result.get(\"pass@1\", \"\u2014\")\nmb = mbpp_result.get(\"pass@1\", \"\u2014\")\nmbp = mbppp_result.get(\"pass@1\", \"\u2014\")\n\nMODEL_CARD = f\"\"\"---\nlicense: apache-2.0\nlanguage: [en]\nbase_model: Qwen/Qwen2.5-Coder-7B-Instruct\ntags: [code, qwen2.5, lora, merged, sft, dpo, grpo]\nlibrary_name: transformers\npipeline_tag: text-generation\nmodel-index:\n- name: TIMPS-Coder-7B\n results:\n - task: {{\"type\": \"text-generation\", \"name\": \"Code Generation\"}}\n dataset: {{\"type\": \"openai_humaneval\", \"name\": \"HumanEval\"}}\n metrics:\n - {{type: \"pass@1\", \"value\": {he}, \"name\": \"pass@1\"}}\n - task: {{\"type\": \"text-generation\", \"name\": \"Code Generation\"}}\n dataset: {{\"type\": \"evalplus/humanevalplus\", \"name\": \"HumanEval+\"}}\n metrics:\n - {{type: \"pass@1\", \"value\": {hep}, \"name\": \"pass@1\"}}\n - task: {{\"type\": \"text-generation\", \"name\": \"Code Generation\"}}\n dataset: {{\"type\": \"mbpp\", \"name\": \"MBPP\"}}\n metrics:\n - {{type: \"pass@1\", \"value\": {mb}, \"name\": \"pass@1\"}}\n - task: {{\"type\": \"text-generation\", \"name\": \"Code Generation\"}}\n dataset: {{\"type\": \"evalplus/mbppplus\", \"name\": \"MBPP+\"}}\n metrics:\n - {{type: \"pass@1\", \"value\": {mbp}, \"name\": \"pass@1\"}}\n---\n\n# TIMPS-Coder-7B\n\nTIMPS-Coder-7B is a code-generation model built by fine-tuning **Qwen2.5-Coder-7B-Instruct**\nthrough a 4-step pipeline: SFT, GRPO, DPO, and Self-Guided Self-Play.\n\n## Benchmark Results\n\n| Benchmark | Score |\n|-----------|-------|\n| **HumanEval pass@1** | **{he}%** |\n| **HumanEval+ pass@1** | **{hep}%** |\n| **MBPP pass@1** | **{mb}%** |\n| **MBPP+ pass@1** | **{mbp}%** |\n\n### Comparison with 7B-9B Code Models\n\n{table}\n\n## Usage\n\n```python\nfrom transformers import AutoModelForCausalLM, AutoTokenizer\nmodel = AutoModelForCausalLM.from_pretrained(\"sandeeprdy1729/TIMPS-Coder-7B\", device_map=\"auto\", torch_dtype=\"auto\")\ntokenizer = AutoTokenizer.from_pretrained(\"sandeeprdy1729/TIMPS-Coder-7B\")\nmessages = [{{\"role\": \"user\", \"content\": \"Write a fibonacci function.\"}}]\ninputs = tokenizer.apply_chat_template(messages, return_tensors=\"pt\").to(model.device)\nprint(tokenizer.decode(model.generate(inputs, max_new_tokens=512)[0]))\n\"\"\"\n\nwith open(f\"{OUTPUT_DIR}/model_card_results.md\", \"w\") as f:\n f.write(MODEL_CARD)\nprint(f\"Model card saved to {OUTPUT_DIR}/model_card_results.md\")\nprint(MODEL_CARD)",
|
||||
"metadata": {
|
||||
"trusted": true
|
||||
},
|
||||
"outputs": [],
|
||||
"execution_count": null
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"source": "\n#**Cell 11 \u2014 Push scores to HuggingFace model card (optional):**\n# ```python\nfrom huggingface_hub import HfApi\n\napi = HfApi()\napi.upload_file(\n path_or_fileobj=f\"{OUTPUT_DIR}/model_card_results.md\",\n path_in_repo=\"README.md\",\n repo_id=MODEL_ID,\n repo_type=\"model\",\n commit_message=f\"Add benchmark results: HE={he}% HE+={hep}% MBPP={mb}% MBPP+={mbp}%\",\n)\nprint(f\"Model card updated at https://huggingface.co/{MODEL_ID}\")",
|
||||
"metadata": {
|
||||
"trusted": true
|
||||
},
|
||||
"outputs": [],
|
||||
"execution_count": null
|
||||
}
|
||||
]
|
||||
}
|
||||
54
chat_template.jinja
Normal file
54
chat_template.jinja
Normal file
@@ -0,0 +1,54 @@
|
||||
{%- if tools %}
|
||||
{{- '<|im_start|>system\n' }}
|
||||
{%- if messages[0]['role'] == 'system' %}
|
||||
{{- messages[0]['content'] }}
|
||||
{%- else %}
|
||||
{{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
|
||||
{%- endif %}
|
||||
{{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
||||
{%- for tool in tools %}
|
||||
{{- "\n" }}
|
||||
{{- tool | tojson }}
|
||||
{%- endfor %}
|
||||
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
|
||||
{%- else %}
|
||||
{%- if messages[0]['role'] == 'system' %}
|
||||
{{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
|
||||
{%- else %}
|
||||
{{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
|
||||
{%- endif %}
|
||||
{%- endif %}
|
||||
{%- for message in messages %}
|
||||
{%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
|
||||
{{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
|
||||
{%- elif message.role == "assistant" %}
|
||||
{{- '<|im_start|>' + message.role }}
|
||||
{%- if message.content %}
|
||||
{{- '\n' + message.content }}
|
||||
{%- endif %}
|
||||
{%- for tool_call in message.tool_calls %}
|
||||
{%- if tool_call.function is defined %}
|
||||
{%- set tool_call = tool_call.function %}
|
||||
{%- endif %}
|
||||
{{- '\n<tool_call>\n{"name": "' }}
|
||||
{{- tool_call.name }}
|
||||
{{- '", "arguments": ' }}
|
||||
{{- tool_call.arguments | tojson }}
|
||||
{{- '}\n</tool_call>' }}
|
||||
{%- endfor %}
|
||||
{{- '<|im_end|>\n' }}
|
||||
{%- elif message.role == "tool" %}
|
||||
{%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
|
||||
{{- '<|im_start|>user' }}
|
||||
{%- endif %}
|
||||
{{- '\n<tool_response>\n' }}
|
||||
{{- message.content }}
|
||||
{{- '\n</tool_response>' }}
|
||||
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
||||
{{- '<|im_end|>\n' }}
|
||||
{%- endif %}
|
||||
{%- endif %}
|
||||
{%- endfor %}
|
||||
{%- if add_generation_prompt %}
|
||||
{{- '<|im_start|>assistant\n' }}
|
||||
{%- endif %}
|
||||
61
config.json
Normal file
61
config.json
Normal file
@@ -0,0 +1,61 @@
|
||||
{
|
||||
"architectures": [
|
||||
"Qwen2ForCausalLM"
|
||||
],
|
||||
"attention_dropout": 0.0,
|
||||
"bos_token_id": 151643,
|
||||
"dtype": "bfloat16",
|
||||
"eos_token_id": 151645,
|
||||
"hidden_act": "silu",
|
||||
"hidden_size": 3584,
|
||||
"initializer_range": 0.02,
|
||||
"intermediate_size": 18944,
|
||||
"layer_types": [
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention",
|
||||
"full_attention"
|
||||
],
|
||||
"max_position_embeddings": 32768,
|
||||
"max_window_layers": 28,
|
||||
"model_type": "qwen2",
|
||||
"num_attention_heads": 28,
|
||||
"num_hidden_layers": 28,
|
||||
"num_key_value_heads": 4,
|
||||
"pad_token_id": null,
|
||||
"rms_norm_eps": 1e-06,
|
||||
"rope_parameters": {
|
||||
"rope_theta": 1000000.0,
|
||||
"rope_type": "default"
|
||||
},
|
||||
"sliding_window": null,
|
||||
"tie_word_embeddings": false,
|
||||
"transformers_version": "5.14.1",
|
||||
"use_cache": true,
|
||||
"use_sliding_window": false,
|
||||
"vocab_size": 152064
|
||||
}
|
||||
3
model-1-of-4.safetensors
Normal file
3
model-1-of-4.safetensors
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:9138a2393e0f690f5f17b6e1fc88fe90c4066014cd8ee5158dea95b029e0b869
|
||||
size 4217008248
|
||||
3
model-2-of-4.safetensors
Normal file
3
model-2-of-4.safetensors
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:a636b7518ff3bcedc4fb69ab159df67cd79cf7d204d623dc85cece8540659e53
|
||||
size 4195052584
|
||||
3
model-3-of-4.safetensors
Normal file
3
model-3-of-4.safetensors
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:55e15f17566a7e2d97ee8ead4e29a6f855c5cadf544c3a3030a011f677178291
|
||||
size 4195052632
|
||||
3
model-4-of-4.safetensors
Normal file
3
model-4-of-4.safetensors
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:90a8911fe69b9eab6572ab085cd08acf8ce5e7ff045f36a1f854970d00c41625
|
||||
size 2624158280
|
||||
346
model.safetensors.index.json
Normal file
346
model.safetensors.index.json
Normal file
@@ -0,0 +1,346 @@
|
||||
{
|
||||
"metadata": {
|
||||
"format": "pt"
|
||||
},
|
||||
"weight_map": {
|
||||
"model.embed_tokens.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.0.self_attn.q_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.0.self_attn.q_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.0.self_attn.k_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.0.self_attn.k_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.0.self_attn.v_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.0.self_attn.v_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.0.self_attn.o_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.0.mlp.gate_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.0.mlp.up_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.0.mlp.down_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.0.input_layernorm.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.0.post_attention_layernorm.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.1.self_attn.q_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.1.self_attn.q_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.1.self_attn.k_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.1.self_attn.k_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.1.self_attn.v_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.1.self_attn.v_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.1.self_attn.o_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.1.mlp.gate_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.1.mlp.up_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.1.mlp.down_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.1.input_layernorm.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.1.post_attention_layernorm.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.2.self_attn.q_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.2.self_attn.q_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.2.self_attn.k_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.2.self_attn.k_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.2.self_attn.v_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.2.self_attn.v_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.2.self_attn.o_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.2.mlp.gate_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.2.mlp.up_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.2.mlp.down_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.2.input_layernorm.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.2.post_attention_layernorm.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.3.self_attn.q_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.3.self_attn.q_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.3.self_attn.k_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.3.self_attn.k_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.3.self_attn.v_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.3.self_attn.v_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.3.self_attn.o_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.3.mlp.gate_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.3.mlp.up_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.3.mlp.down_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.3.input_layernorm.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.3.post_attention_layernorm.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.4.self_attn.q_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.4.self_attn.q_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.4.self_attn.k_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.4.self_attn.k_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.4.self_attn.v_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.4.self_attn.v_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.4.self_attn.o_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.4.mlp.gate_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.4.mlp.up_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.4.mlp.down_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.4.input_layernorm.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.4.post_attention_layernorm.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.5.self_attn.q_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.5.self_attn.q_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.5.self_attn.k_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.5.self_attn.k_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.5.self_attn.v_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.5.self_attn.v_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.5.self_attn.o_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.5.mlp.gate_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.5.mlp.up_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.5.mlp.down_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.5.input_layernorm.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.5.post_attention_layernorm.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.6.self_attn.q_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.6.self_attn.q_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.6.self_attn.k_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.6.self_attn.k_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.6.self_attn.v_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.6.self_attn.v_proj.bias": "model-1-of-4.safetensors",
|
||||
"model.layers.6.self_attn.o_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.6.mlp.gate_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.6.mlp.up_proj.weight": "model-1-of-4.safetensors",
|
||||
"model.layers.6.mlp.down_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.6.input_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.6.post_attention_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.7.self_attn.q_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.7.self_attn.q_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.7.self_attn.k_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.7.self_attn.k_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.7.self_attn.v_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.7.self_attn.v_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.7.self_attn.o_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.7.mlp.gate_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.7.mlp.up_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.7.mlp.down_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.7.input_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.7.post_attention_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.8.self_attn.q_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.8.self_attn.q_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.8.self_attn.k_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.8.self_attn.k_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.8.self_attn.v_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.8.self_attn.v_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.8.self_attn.o_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.8.mlp.gate_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.8.mlp.up_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.8.mlp.down_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.8.input_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.8.post_attention_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.9.self_attn.q_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.9.self_attn.q_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.9.self_attn.k_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.9.self_attn.k_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.9.self_attn.v_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.9.self_attn.v_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.9.self_attn.o_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.9.mlp.gate_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.9.mlp.up_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.9.mlp.down_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.9.input_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.9.post_attention_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.10.self_attn.q_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.10.self_attn.q_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.10.self_attn.k_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.10.self_attn.k_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.10.self_attn.v_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.10.self_attn.v_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.10.self_attn.o_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.10.mlp.gate_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.10.mlp.up_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.10.mlp.down_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.10.input_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.10.post_attention_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.11.self_attn.q_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.11.self_attn.q_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.11.self_attn.k_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.11.self_attn.k_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.11.self_attn.v_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.11.self_attn.v_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.11.self_attn.o_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.11.mlp.gate_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.11.mlp.up_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.11.mlp.down_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.11.input_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.11.post_attention_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.12.self_attn.q_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.12.self_attn.q_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.12.self_attn.k_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.12.self_attn.k_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.12.self_attn.v_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.12.self_attn.v_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.12.self_attn.o_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.12.mlp.gate_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.12.mlp.up_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.12.mlp.down_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.12.input_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.12.post_attention_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.13.self_attn.q_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.13.self_attn.q_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.13.self_attn.k_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.13.self_attn.k_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.13.self_attn.v_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.13.self_attn.v_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.13.self_attn.o_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.13.mlp.gate_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.13.mlp.up_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.13.mlp.down_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.13.input_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.13.post_attention_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.14.self_attn.q_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.14.self_attn.q_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.14.self_attn.k_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.14.self_attn.k_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.14.self_attn.v_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.14.self_attn.v_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.14.self_attn.o_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.14.mlp.gate_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.14.mlp.up_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.14.mlp.down_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.14.input_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.14.post_attention_layernorm.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.15.self_attn.q_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.15.self_attn.q_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.15.self_attn.k_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.15.self_attn.k_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.15.self_attn.v_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.15.self_attn.v_proj.bias": "model-2-of-4.safetensors",
|
||||
"model.layers.15.self_attn.o_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.15.mlp.gate_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.15.mlp.up_proj.weight": "model-2-of-4.safetensors",
|
||||
"model.layers.15.mlp.down_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.15.input_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.15.post_attention_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.16.self_attn.q_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.16.self_attn.q_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.16.self_attn.k_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.16.self_attn.k_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.16.self_attn.v_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.16.self_attn.v_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.16.self_attn.o_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.16.mlp.gate_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.16.mlp.up_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.16.mlp.down_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.16.input_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.16.post_attention_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.17.self_attn.q_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.17.self_attn.q_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.17.self_attn.k_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.17.self_attn.k_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.17.self_attn.v_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.17.self_attn.v_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.17.self_attn.o_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.17.mlp.gate_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.17.mlp.up_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.17.mlp.down_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.17.input_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.17.post_attention_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.18.self_attn.q_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.18.self_attn.q_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.18.self_attn.k_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.18.self_attn.k_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.18.self_attn.v_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.18.self_attn.v_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.18.self_attn.o_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.18.mlp.gate_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.18.mlp.up_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.18.mlp.down_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.18.input_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.18.post_attention_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.19.self_attn.q_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.19.self_attn.q_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.19.self_attn.k_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.19.self_attn.k_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.19.self_attn.v_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.19.self_attn.v_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.19.self_attn.o_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.19.mlp.gate_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.19.mlp.up_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.19.mlp.down_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.19.input_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.19.post_attention_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.20.self_attn.q_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.20.self_attn.q_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.20.self_attn.k_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.20.self_attn.k_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.20.self_attn.v_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.20.self_attn.v_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.20.self_attn.o_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.20.mlp.gate_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.20.mlp.up_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.20.mlp.down_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.20.input_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.20.post_attention_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.21.self_attn.q_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.21.self_attn.q_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.21.self_attn.k_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.21.self_attn.k_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.21.self_attn.v_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.21.self_attn.v_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.21.self_attn.o_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.21.mlp.gate_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.21.mlp.up_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.21.mlp.down_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.21.input_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.21.post_attention_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.22.self_attn.q_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.22.self_attn.q_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.22.self_attn.k_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.22.self_attn.k_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.22.self_attn.v_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.22.self_attn.v_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.22.self_attn.o_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.22.mlp.gate_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.22.mlp.up_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.22.mlp.down_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.22.input_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.22.post_attention_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.23.self_attn.q_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.23.self_attn.q_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.23.self_attn.k_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.23.self_attn.k_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.23.self_attn.v_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.23.self_attn.v_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.23.self_attn.o_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.23.mlp.gate_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.23.mlp.up_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.23.mlp.down_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.23.input_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.23.post_attention_layernorm.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.24.self_attn.q_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.24.self_attn.q_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.24.self_attn.k_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.24.self_attn.k_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.24.self_attn.v_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.24.self_attn.v_proj.bias": "model-3-of-4.safetensors",
|
||||
"model.layers.24.self_attn.o_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.24.mlp.gate_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.24.mlp.up_proj.weight": "model-3-of-4.safetensors",
|
||||
"model.layers.24.mlp.down_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.24.input_layernorm.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.24.post_attention_layernorm.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.25.self_attn.q_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.25.self_attn.q_proj.bias": "model-4-of-4.safetensors",
|
||||
"model.layers.25.self_attn.k_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.25.self_attn.k_proj.bias": "model-4-of-4.safetensors",
|
||||
"model.layers.25.self_attn.v_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.25.self_attn.v_proj.bias": "model-4-of-4.safetensors",
|
||||
"model.layers.25.self_attn.o_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.25.mlp.gate_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.25.mlp.up_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.25.mlp.down_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.25.input_layernorm.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.25.post_attention_layernorm.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.26.self_attn.q_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.26.self_attn.q_proj.bias": "model-4-of-4.safetensors",
|
||||
"model.layers.26.self_attn.k_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.26.self_attn.k_proj.bias": "model-4-of-4.safetensors",
|
||||
"model.layers.26.self_attn.v_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.26.self_attn.v_proj.bias": "model-4-of-4.safetensors",
|
||||
"model.layers.26.self_attn.o_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.26.mlp.gate_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.26.mlp.up_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.26.mlp.down_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.26.input_layernorm.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.26.post_attention_layernorm.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.27.self_attn.q_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.27.self_attn.q_proj.bias": "model-4-of-4.safetensors",
|
||||
"model.layers.27.self_attn.k_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.27.self_attn.k_proj.bias": "model-4-of-4.safetensors",
|
||||
"model.layers.27.self_attn.v_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.27.self_attn.v_proj.bias": "model-4-of-4.safetensors",
|
||||
"model.layers.27.self_attn.o_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.27.mlp.gate_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.27.mlp.up_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.27.mlp.down_proj.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.27.input_layernorm.weight": "model-4-of-4.safetensors",
|
||||
"model.layers.27.post_attention_layernorm.weight": "model-4-of-4.safetensors",
|
||||
"model.norm.weight": "model-4-of-4.safetensors",
|
||||
"lm_head.weight": "model-4-of-4.safetensors"
|
||||
}
|
||||
}
|
||||
3
tokenizer.json
Normal file
3
tokenizer.json
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8
|
||||
size 11421892
|
||||
30
tokenizer_config.json
Normal file
30
tokenizer_config.json
Normal file
@@ -0,0 +1,30 @@
|
||||
{
|
||||
"add_prefix_space": false,
|
||||
"backend": "tokenizers",
|
||||
"bos_token": null,
|
||||
"clean_up_tokenization_spaces": false,
|
||||
"eos_token": "<|im_end|>",
|
||||
"errors": "replace",
|
||||
"extra_special_tokens": [
|
||||
"<|im_start|>",
|
||||
"<|im_end|>",
|
||||
"<|object_ref_start|>",
|
||||
"<|object_ref_end|>",
|
||||
"<|box_start|>",
|
||||
"<|box_end|>",
|
||||
"<|quad_start|>",
|
||||
"<|quad_end|>",
|
||||
"<|vision_start|>",
|
||||
"<|vision_end|>",
|
||||
"<|vision_pad|>",
|
||||
"<|image_pad|>",
|
||||
"<|video_pad|>"
|
||||
],
|
||||
"is_local": false,
|
||||
"local_files_only": false,
|
||||
"model_max_length": 32768,
|
||||
"pad_token": "<|endoftext|>",
|
||||
"split_special_tokens": false,
|
||||
"tokenizer_class": "Qwen2Tokenizer",
|
||||
"unk_token": null
|
||||
}
|
||||
Reference in New Issue
Block a user