File size: 20,783 Bytes
1efeea2 bbd90f1 1efeea2 dd9ce5a 1efeea2 dd9ce5a 1efeea2 dd9ce5a 1efeea2 408b6e9 1efeea2 408b6e9 1efeea2 f6d0ad7 dd9ce5a 1efeea2 dd9ce5a 1efeea2 96ef4b5 1efeea2 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 |
---
pipeline_tag: text-generation
inference: false
license: apache-2.0
library_name: transformers
tags:
- language
- granite-3.1
---
# Granite-3.1-1B-A400M-Base
**Model Summary:**
Granite-3.1-1B-A400M-Base extends the context length of Granite-3.0-1B-A400M-Base from 4K to 128K using a progressive training strategy by increasing the supported context length in increments while adjusting RoPE theta until the model has successfully adapted to desired length of 128K. This long-context pre-training stage was performed using approximately 500B tokens.
- **Developers:** Granite Team, IBM
- **GitHub Repository:** [ibm-granite/granite-3.1-language-models](https://github.com/ibm-granite/granite-3.1-language-models)
- **Website**: [Granite Docs](https://www.ibm.com/granite/docs/)
- **Paper:** [Granite 3.1 Language Models (coming soon)](https://huggingface.co/collections/ibm-granite/granite-31-language-models-6751dbbf2f3389bec5c6f02d)
- **Release Date**: December 18th, 2024
- **License:** [Apache 2.0](https://www.apache.org/licenses/LICENSE-2.0)
**Supported Languages:**
English, German, Spanish, French, Japanese, Portuguese, Arabic, Czech, Italian, Korean, Dutch, and Chinese. Users may finetune Granite 3.1 models for languages beyond these 12 languages.
**Intended Use:**
Prominent use cases of LLMs in text-to-text generation include summarization, text classification, extraction, question-answering, and more. All Granite Base models are able to handle these tasks as they were trained on a large amount of data from various domains. Moreover, they can serve as baseline to create specialized models for specific application scenarios.
**Generation:**
This is a simple example of how to use Granite-3.1-1B-A400M-Base model.
Install the following libraries:
```shell
pip install torch torchvision torchaudio
pip install accelerate
pip install transformers
```
Then, copy the code snippet below to run the example.
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
device = "auto"
model_path = "ibm-granite/granite-3.1-1b-a400m-base"
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 = "Where is the Thomas J. Watson Research Center located?"
# tokenize the text
input_tokens = tokenizer(input_text, return_tensors="pt").to(device)
# generate output tokens
output = model.generate(**input_tokens,
max_length=4000)
# decode output tokens into text
output = tokenizer.batch_decode(output)
# print output
print(output)
```
**Evaluation Results:**
<table>
<caption><b>HuggingFace Open LLM Leaderboard V1</b></caption>
<thead>
<tr>
<th style="text-align:left; background-color: #001d6c; color: white;">Models</th>
<th style="text-align:center; background-color: #001d6c; color: white;">ARC-Challenge</th>
<th style="text-align:center; background-color: #001d6c; color: white;">Hellaswag</th>
<th style="text-align:center; background-color: #001d6c; color: white;">MMLU</th>
<th style="text-align:center; background-color: #001d6c; color: white;">TruthfulQA</th>
<th style="text-align:center; background-color: #001d6c; color: white;">Winogrande</th>
<th style="text-align:center; background-color: #001d6c; color: white;">GSM8K</th>
<th style="text-align:center; background-color: #001d6c; color: white;">Avg</th>
</tr></thead>
<tbody>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: black;">Granite-3.1-8B-Base</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">63.99</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">83.27</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">63.45</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">51.29</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">78.92</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">60.19</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">66.85</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">Granite-3.1-2B-Base</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">53.58</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">77.67</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">52.86</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">39.02</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">72.84</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">47.99</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">57.32</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">Granite-3.1-3B-A800M-Base</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">50.76</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">74.45</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">48.31</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">39.91</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">69.29</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">40.56</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">53.88</td>
</tr>
<tr>
<td style="text-align:left; background-color: #DAE8FF; color: #2D2D2D;">Granite-3.1-1B-A400M-Base</td>
<td style="text-align:center; background-color: #DAE8FF; color: #2D2D2D;">39.42</td>
<td style="text-align:center; background-color: #DAE8FF; color: #2D2D2D;">66.13</td>
<td style="text-align:center; background-color: #DAE8FF; color: #2D2D2D;">26.53</td>
<td style="text-align:center; background-color: #DAE8FF; color: #2D2D2D;">37.67</td>
<td style="text-align:center; background-color: #DAE8FF; color: #2D2D2D;">2.03</td>
<td style="text-align:center; background-color: #DAE8FF; color: #2D2D2D;">18.87</td>
<td style="text-align:center; background-color: #DAE8FF; color: #2D2D2D;">31.78</td>
</tr>
</tbody></table>
<table>
<caption><b>HuggingFace Open LLM Leaderboard V2</b></caption>
<thead>
<tr>
<th style="text-align:left; background-color: #001d6c; color: white;">Models</th>
<th style="text-align:center; background-color: #001d6c; color: white;">IFEval</th>
<th style="text-align:center; background-color: #001d6c; color: white;">BBH</th>
<th style="text-align:center; background-color: #001d6c; color: white;">MATH Lvl 5</th>
<th style="text-align:center; background-color: #001d6c; color: white;">GPQA</th>
<th style="text-align:center; background-color: #001d6c; color: white;">MUSR</th>
<th style="text-align:center; background-color: #001d6c; color: white;">MMLU-Pro</th>
<th style="text-align:center; background-color: #001d6c; color: white;">Avg</th>
</tr></thead>
<tbody>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: black;">Granite-3.1-8B-Base</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">42.21</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">26.02</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">9.52</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">9.51</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">8.36</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">24.8</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">20.07</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">Granite-3.1-2B-Base</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">35.22</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">16.84</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">5.59</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">3.69</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">3.9</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">13.9</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">13.19</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">Granite-3.1-3B-A800M-Base</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">29.96</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">11.91</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">4</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">3.69</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">1.11</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">8.81</td>
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">9.91</td>
</tr>
<tr>
<td style="text-align:left; background-color: #DAE8FF; color: #2D2D2D;">Granite-3.1-1B-A400M-Base</td>
<td style="text-align:center; background-color: #DAE8FF; color: #2D2D2D;">25.19</td>
<td style="text-align:center; background-color: #DAE8FF; color: #2D2D2D;">6.43</td>
<td style="text-align:center; background-color: #DAE8FF; color: #2D2D2D;">2.19</td>
<td style="text-align:center; background-color: #DAE8FF; color: #2D2D2D;">0.22</td>
<td style="text-align:center; background-color: #DAE8FF; color: #2D2D2D;">1.76</td>
<td style="text-align:center; background-color: #DAE8FF; color: #2D2D2D;">1.55</td>
<td style="text-align:center; background-color: #DAE8FF; color: #2D2D2D;">6.22</td>
</tr>
</tbody></table>
**Model Architecture:**
Granite-3.1-1B-A400M-Base is based on a decoder-only sparse Mixture of Experts (MoE) transformer architecture. Core components of this architecture are: Fine-grained Experts, Dropless Token Routing, and Load Balancing Loss.
<table>
<thead>
<tr>
<th style="text-align:left; background-color: #001d6c; color: white;">Model</th>
<th style="text-align:center; background-color: #001d6c; color: white;">2B Dense</th>
<th style="text-align:center; background-color: #001d6c; color: white;">8B Dense</th>
<th style="text-align:center; background-color: #001d6c; color: white;">1B MoE</th>
<th style="text-align:center; background-color: #001d6c; color: white;">3B MoE</th>
</tr></thead>
<tbody>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: black;">Embedding size</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">2048</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">4096</td>
<td style="text-align:center; background-color: #DAE8FF; color: black;">1024</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">1536</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: black;">Number of layers</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">40</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">40</td>
<td style="text-align:center; background-color: #DAE8FF; color: black;">24</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">32</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: black;">Attention head size</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">64</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">128</td>
<td style="text-align:center; background-color: #DAE8FF; color: black;">64</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">64</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: black;">Number of attention heads</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">32</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">32</td>
<td style="text-align:center; background-color: #DAE8FF; color: black;">16</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">24</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: black;">Number of KV heads</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">8</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">8</td>
<td style="text-align:center; background-color: #DAE8FF; color: black;">8</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">8</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: black;">MLP hidden size</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">8192</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">12800</td>
<td style="text-align:center; background-color: #DAE8FF; color: black;">512</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">512</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: black;">MLP activation</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">SwiGLU</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">SwiGLU</td>
<td style="text-align:center; background-color: #DAE8FF; color: black;">SwiGLU</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">SwiGLU</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: black;">Number of experts</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">—</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">—</td>
<td style="text-align:center; background-color: #DAE8FF; color: black;">32</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">40</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: black;">MoE TopK</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">—</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">—</td>
<td style="text-align:center; background-color: #DAE8FF; color: black;">8</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">8</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: black;">Initialization std</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">0.1</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">0.1</td>
<td style="text-align:center; background-color: #DAE8FF; color: black;">0.1</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">0.1</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: black;">Sequence length</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">128K</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">128K</td>
<td style="text-align:center; background-color: #DAE8FF; color: black;">128K</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">128K</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: black;">Position embedding</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">RoPE</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">RoPE</td>
<td style="text-align:center; background-color: #DAE8FF; color: black;">RoPE</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">RoPE</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: black;"># Parameters</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">2.5B</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">8.1B</td>
<td style="text-align:center; background-color: #DAE8FF; color: black;">1.3B</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">3.3B</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: black;"># Active parameters</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">2.5B</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">8.1B</td>
<td style="text-align:center; background-color: #DAE8FF; color: black;">400M</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">800M</td>
</tr>
<tr>
<td style="text-align:left; background-color: #FFFFFF; color: black;"># Training tokens</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">12T</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">12T</td>
<td style="text-align:center; background-color: #DAE8FF; color: black;">10T</td>
<td style="text-align:center; background-color: #FFFFFF; color: black;">10T</td>
</tr>
</tbody></table>
**Training Data:**
This model is trained on a mix of open source and proprietary data following a two-stage training strategy.
* Stage 1 data: The data for stage 1 is sourced from diverse domains, such as: web, code, academic sources, books, and math data.
* Stage 2 data: The data for stage 2 comprises a curated mix of high-quality data from the same domains, plus multilingual and instruction data. The goal of this second training phase is to enhance the model’s performance on specific tasks.
* Stage 3 data: The data for stage 3 consists of original stage-2 pretraining data with additional synthetic long-context data in form of QA/summary pairs where the answer contains a recitation of the related paragraph before the answer.
A detailed attribution of datasets can be found in the [Granite 3.0 Technical Report](https://github.com/ibm-granite/granite-3.0-language-models/blob/main/paper.pdf), [Granite 3.1 Technical Report (coming soon)](https://huggingface.co/collections/ibm-granite/granite-31-language-models-6751dbbf2f3389bec5c6f02d), and [Accompanying Author List](https://github.com/ibm-granite/granite-3.0-language-models/blob/main/author-ack.pdf).
**Infrastructure:**
We train Granite 3.1 Language Models using IBM's super computing cluster, Blue Vela, which is outfitted with NVIDIA H100 GPUs. This cluster provides 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, including but not limited to: bias and fairness, misinformation, and autonomous decision-making. Granite-3.1-1B-A400M-Base model is not the exception in this regard. Even though this model is suited for multiple generative AI 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 text 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-3.1-1B-A400M-Base model with ethical intentions and in a responsible way.
**Resources**
- ⭐️ Learn about the latest updates with Granite: https://www.ibm.com/granite
- 📄 Get started with tutorials, best practices, and prompt engineering advice: https://www.ibm.com/granite/docs/
- 💡 Learn about the latest Granite learning resources: https://ibm.biz/granite-learning-resources
<!-- ## Citation
```
@misc{granite-models,
author = {author 1, author2, ...},
title = {},
journal = {},
volume = {},
year = {2024},
url = {https://arxiv.org/abs/0000.00000},
}
``` -->
|