File size: 9,002 Bytes
9a610a1 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 |
Quantization made by Richard Erkhov.
[Github](https://github.com/RichardErkhov)
[Discord](https://discord.gg/pvy7H8DZMG)
[Request more models](https://github.com/RichardErkhov/quant_request)
granite-3b-code-base-128k - AWQ
- Model creator: https://huggingface.co/ibm-granite/
- Original model: https://huggingface.co/ibm-granite/granite-3b-code-base-128k/
Original model description:
---
pipeline_tag: text-generation
inference: false
license: apache-2.0
datasets:
- codeparrot/github-code-clean
- bigcode/starcoderdata
# - Stackexchange
# - CommonCrawl
- open-web-math/open-web-math
- math-ai/StackMathQA
# - Arxiv
# - Wikipedia
# - conceptofmind/FLAN_2022 # Original link is broken, we used IBM's filtered version
metrics:
- code_eval
library_name: transformers
tags:
- code
- granite
model-index:
- name: granite-3b-code-base-128k
results:
- task:
type: text-generation
dataset:
type: bigcode/humanevalpack
name: HumanEvalSynthesis (Python)
metrics:
- name: pass@1
type: pass@1
value: 36.0
verified: false
- task:
type: text-generation
dataset:
type: bigcode/humanevalpack
name: HumanEvalSynthesis (Average)
metrics:
- name: pass@1
type: pass@1
value: 30.5
verified: false
- task:
type: text-generation
dataset:
type: bigcode/humanevalpack
name: HumanEvalExplain (Average)
metrics:
- name: pass@1
type: pass@1
value: 22.4
verified: false
- task:
type: text-generation
dataset:
type: bigcode/humanevalpack
name: HumanEvalFix (Average)
metrics:
- name: pass@1
type: pass@1
value: 19.9
verified: false
- task:
type: text-generation
dataset:
type: repoqa
name: RepoQA (Python@16K)
metrics:
- name: pass@1 (thresh=0.5)
type: pass@1 (thresh=0.5)
value: 40.0
verified: false
- task:
type: text-generation
dataset:
type: repoqa
name: RepoQA (C++@16K)
metrics:
- name: pass@1 (thresh=0.5)
type: pass@1 (thresh=0.5)
value: 36.0
verified: false
- task:
type: text-generation
dataset:
type: repoqa
name: RepoQA (Java@16K)
metrics:
- name: pass@1 (thresh=0.5)
type: pass@1 (thresh=0.5)
value: 37.0
verified: false
- task:
type: text-generation
dataset:
type: repoqa
name: RepoQA (TypeScript@16K)
metrics:
- name: pass@1 (thresh=0.5)
type: pass@1 (thresh=0.5)
value: 27.0
verified: false
- task:
type: text-generation
dataset:
type: repoqa
name: RepoQA (Rust@16K)
metrics:
- name: pass@1 (thresh=0.5)
type: pass@1 (thresh=0.5)
value: 29.0
verified: false
- task:
type: text-generation
dataset:
type: lcc
name: LCC (Balanced)
metrics:
- name: Exact Match@4K
type: Exact Match@4K
value: 54.6
verified: false
- task:
type: text-generation
dataset:
type: lcc
name: LCC (Balanced)
metrics:
- name: Exact Match@8K
type: Exact Match@8K
value: 56.8
verified: false
- task:
type: text-generation
dataset:
type: lcc
name: LCC (Balanced)
metrics:
- name: Exact Match@16K
type: Exact Match@16K
value: 52.2
verified: false
- task:
type: text-generation
dataset:
type: lcc
name: LCC (Balanced)
metrics:
- name: Exact Match@32K
type: Exact Match@32K
value: 57.8
verified: false
- task:
type: text-generation
dataset:
type: repobench
name: RepoBench-P (Balanced)
metrics:
- name: Exact Match@4K
type: Exact Match@4K
value: 39.8
verified: false
- task:
type: text-generation
dataset:
type: repobench
name: RepoBench-P (Balanced)
metrics:
- name: Exact Match@8K
type: Exact Match@8K
value: 46.8
verified: false
- task:
type: text-generation
dataset:
type: repobench
name: RepoBench-P (Balanced)
metrics:
- name: Exact Match@16K
type: Exact Match@16K
value: 43.1
verified: false
- task:
type: text-generation
dataset:
type: repobench
name: RepoBench-Pn(Balanced)
metrics:
- name: Exact Match@32K
type: Exact Match@32K
value: 45.3
verified: false
---
# Granite-3B-Code-Base-128K
## Model Summary
**Granite-3B-Code-Base-128K** extends the context length of Granite-3B-Code-Base from 2K to 128K with continual pretraining using the original training data but with repository-level file packing and per-language length upsampling, that we found to be critical for long-context pretraining.
We adopt an progressive training strategy where we doubled the context window until it reached the desired length of 128K by appropriately adjusting RoPE theta. We trained on 4B tokens total for all stages, which is only 0.1% of Granite-3B-Code-Base's original pre-training data.
- **Developers:** IBM Research
- **GitHub Repository:** [ibm-granite/granite-code-models](https://github.com/ibm-granite/granite-code-models)
- **Paper:** [Scaling Granite Code Models to 128K Context](https://arxiv.org/abs/2405.04324)
- **Release Date**: July 18th, 2024
- **License:** [Apache 2.0](https://www.apache.org/licenses/LICENSE-2.0).
## Usage
### Intended use
Prominent enterprise use cases of LLMs in software engineering productivity with 128K context length support that includes code generation, code explanation, code fixing, generating unit tests, generating documentation, addressing technical debt issues, vulnerability detection, code translation, and more. All Granite Code Base models, including the **3B parameter model**, are able to handle these tasks as they were trained on a large amount of code data from 116 programming languages.
### Generation
This is a simple example of how to use **Granite-3B-Code-Base-128K** model.
```python
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
device = "cuda" # or "cpu"
model_path = "ibm-granite/granite-3b-code-base-128k"
tokenizer = AutoTokenizer.from_pretrained(model_path)
# drop device_map if running on CPU
model = AutoModelForCausalLM.from_pretrained(model_path, device_map=device)
model.eval()
# change input text as desired
input_text = "def generate():"
# tokenize the text
input_tokens = tokenizer(input_text, return_tensors="pt")
# transfer tokenized inputs to the device
for i in input_tokens:
input_tokens[i] = input_tokens[i].to(device)
# generate output tokens
output = model.generate(**input_tokens)
# decode output tokens into text
output = tokenizer.batch_decode(output)
# loop over the batch to print, in this example the batch size is 1
for i in output:
print(i)
```
## Training Data
Starting from the base Granite model, this model was further pretrained on repository-level code data with per-language context-length oversampling, allowing it to effectively utilize up to 128K tokens of context. This continued training stage focused on a curated selection of programming languages, such as Python, C, C++, Go, Java, JavaScript, and TypeScript.
## Infrastructure
We train the Granite Code models using two of IBM's super computing clusters, namely Vela and Blue Vela, both outfitted with NVIDIA A100 and H100 GPUs respectively. These clusters provide a scalable and efficient infrastructure for training our models over thousands of GPUs.
## Ethical Considerations and Limitations
The use of Large Language Models involves risks and ethical considerations people must be aware of. Regarding code generation, caution is urged against complete reliance on specific code models for crucial decisions or impactful information as the generated code is not guaranteed to work as intended. **Granite-3B-Code-Base-128K** model is not the exception in this regard. Even though this model is suited for multiple code-related tasks, it has not undergone any safety alignment, there it may produce problematic outputs. Additionally, it remains uncertain whether smaller models might exhibit increased susceptibility to hallucination in generation scenarios by copying source code verbatim from the training dataset due to their reduced sizes and memorization capacities. This aspect is currently an active area of research, and we anticipate more rigorous exploration, comprehension, and mitigations in this domain. Regarding ethics, a latent risk associated with all Large Language Models is their malicious utilization. We urge the community to use **Granite-3B-Code-Base-128K** model with ethical intentions and in a responsible way.
|