model card upload
Browse files
README.md
CHANGED
|
@@ -1,3 +1,435 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
library_name: transformers
|
| 4 |
+
tags:
|
| 5 |
+
- language
|
| 6 |
+
- granite-4.0
|
| 7 |
+
---
|
| 8 |
+
|
| 9 |
+
# Granite-4.0-H-300M-Base
|
| 10 |
+
|
| 11 |
+
**Model Summary:**
|
| 12 |
+
Granite-4.0-H-300M-Base is a lightweight decoder-only language model designed for scenarios where efficiency and speed are critical. They can run on resource-constrained devices such as smartphones or IoT hardware, enabling offline and privacy-preserving applications. It also supports Fill-in-the-Middle (FIM) code completion through the use of specialized prefix and suffix tokens. The model is trained from scratch on approximately 15 trillion tokens following a four-stage training strategy: 10 trillion tokens in the first stage, 2 trillion in the second, another 2 trillion in the third, and 0.5 trillion in the final stage.
|
| 13 |
+
|
| 14 |
+
- **Developers:** Granite Team, IBM
|
| 15 |
+
- **HF Collection:** [Granite 4.0 Nano Language Models HF Collection](https://huggingface.co/collections/ibm-granite/granite-40-nano-language-models-68e5775c80b60e43b72cfa16)
|
| 16 |
+
- **GitHub Repository:** [ibm-granite/granite-4.0-nano-language-models](https://github.com/ibm-granite/granite-4.0-nano-language-models)
|
| 17 |
+
- **Website**: [Granite Docs](https://www.ibm.com/granite/docs/)
|
| 18 |
+
- **Release Date**: October 23, 2025
|
| 19 |
+
- **License:** [Apache 2.0](https://www.apache.org/licenses/LICENSE-2.0)
|
| 20 |
+
|
| 21 |
+
**Supported Languages:**
|
| 22 |
+
English, German, Spanish, French, Japanese, Portuguese, Arabic, Czech, Italian, Korean, Dutch, and Chinese. Users may fine-tune Granite 4.0 Nano models to support languages beyond those included in this list.
|
| 23 |
+
|
| 24 |
+
**Intended Use:**
|
| 25 |
+
Prominent use cases of LLMs in text-to-text generation include summarization, text classification, extraction, question-answering and code-completion (including FIM) tasks. Moreover, these lightweight models can serve as baseline to create task-sepcific models for different applications.
|
| 26 |
+
|
| 27 |
+
**Generation:**
|
| 28 |
+
This is a simple example of how to use Granite-4.0-H-300M-Base model.
|
| 29 |
+
|
| 30 |
+
Install the following libraries:
|
| 31 |
+
|
| 32 |
+
```shell
|
| 33 |
+
pip install torch torchvision torchaudio
|
| 34 |
+
pip install accelerate
|
| 35 |
+
pip install transformers
|
| 36 |
+
```
|
| 37 |
+
Then, copy the code snippet below to run the example.
|
| 38 |
+
|
| 39 |
+
```python
|
| 40 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
| 41 |
+
device = "cuda"
|
| 42 |
+
|
| 43 |
+
model_path = "ibm-granite/granite-4.0-h-300m-base"
|
| 44 |
+
|
| 45 |
+
tokenizer = AutoTokenizer.from_pretrained(model_path)
|
| 46 |
+
# drop device_map if running on CPU
|
| 47 |
+
model = AutoModelForCausalLM.from_pretrained(model_path, device_map=device)
|
| 48 |
+
model.eval()
|
| 49 |
+
# change input text as desired
|
| 50 |
+
input_text = "The capital of France is"
|
| 51 |
+
# tokenize the text
|
| 52 |
+
input_tokens = tokenizer(input_text, return_tensors="pt").to(device)
|
| 53 |
+
# generate output tokens
|
| 54 |
+
output = model.generate(**input_tokens, max_length=10)
|
| 55 |
+
# decode output tokens into text
|
| 56 |
+
output = tokenizer.batch_decode(output)
|
| 57 |
+
# print output
|
| 58 |
+
print(output[0])
|
| 59 |
+
```
|
| 60 |
+
|
| 61 |
+
Expected output:
|
| 62 |
+
```shell
|
| 63 |
+
The capital of France is Paris.
|
| 64 |
+
```
|
| 65 |
+
|
| 66 |
+
**Evaluation Results:**
|
| 67 |
+
<table>
|
| 68 |
+
<thead>
|
| 69 |
+
<tr>
|
| 70 |
+
<th style="text-align:left; background-color: #001d6c; color: white;">Benchmarks</th>
|
| 71 |
+
<th style="text-align:left; background-color: #001d6c; color: white;">Metric</th>
|
| 72 |
+
<th style="text-align:center; background-color: #001d6c; color: white;">300M Dense</th>
|
| 73 |
+
<th style="text-align:center; background-color: #001d6c; color: white;">H 300M Dense</th>
|
| 74 |
+
<th style="text-align:center; background-color: #001d6c; color: white;">1B Dense</th>
|
| 75 |
+
<th style="text-align:center; background-color: #001d6c; color: white;">H 1B Dense</th>
|
| 76 |
+
</tr>
|
| 77 |
+
</thead>
|
| 78 |
+
<tbody>
|
| 79 |
+
<tr>
|
| 80 |
+
<td colspan="6" style="text-align:center; background-color: #FFFFFF; color: #2D2D2D; font-style:italic;">
|
| 81 |
+
General Tasks
|
| 82 |
+
</td>
|
| 83 |
+
</tr>
|
| 84 |
+
<tr>
|
| 85 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">MMLU</td>
|
| 86 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">5-shot</td>
|
| 87 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">33.08</td>
|
| 88 |
+
<td style="text-align:right; background-color: #DAE8FF; color: #2D2D2D;">36.07</td>
|
| 89 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">59.82</td>
|
| 90 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">58.71</td>
|
| 91 |
+
</tr>
|
| 92 |
+
<tr>
|
| 93 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">MMLU-Pro</td>
|
| 94 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">5-shot,CoT</td>
|
| 95 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">11.29</td>
|
| 96 |
+
<td style="text-align:right; background-color: #DAE8FF; color: #2D2D2D;">10.08</td>
|
| 97 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">29.96</td>
|
| 98 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">23.45</td>
|
| 99 |
+
</tr>
|
| 100 |
+
<tr>
|
| 101 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">BBH</td>
|
| 102 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">3-shot, CoT</td>
|
| 103 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">32.19</td>
|
| 104 |
+
<td style="text-align:right; background-color: #DAE8FF; color: #2D2D2D;">29.96</td>
|
| 105 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">57.73</td>
|
| 106 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">48.45</td>
|
| 107 |
+
</tr>
|
| 108 |
+
<tr>
|
| 109 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">AGI EVAL</td>
|
| 110 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">3-shot</td>
|
| 111 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">28.97</td>
|
| 112 |
+
<td style="text-align:right; background-color: #DAE8FF; color: #2D2D2D;">29.2</td>
|
| 113 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">48.95</td>
|
| 114 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">47.46</td>
|
| 115 |
+
</tr>
|
| 116 |
+
<tr>
|
| 117 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">DROP</td>
|
| 118 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">5-shot</td>
|
| 119 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">29.77</td>
|
| 120 |
+
<td style="text-align:right; background-color: #DAE8FF; color: #2D2D2D;">28.56</td>
|
| 121 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">58.18</td>
|
| 122 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">57.18</td>
|
| 123 |
+
</tr>
|
| 124 |
+
<tr>
|
| 125 |
+
<td colspan="6" style="text-align:center; background-color: #FFFFFF; color: #2D2D2D; font-style:italic;">
|
| 126 |
+
Math Tasks
|
| 127 |
+
</td>
|
| 128 |
+
</tr>
|
| 129 |
+
<tr>
|
| 130 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">GSM8K</td>
|
| 131 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">8-shot</td>
|
| 132 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">24.11</td>
|
| 133 |
+
<td style="text-align:right; background-color: #DAE8FF; color: #2D2D2D;">24.41</td>
|
| 134 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">62.4</td>
|
| 135 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">57.39</td>
|
| 136 |
+
</tr>
|
| 137 |
+
<tr>
|
| 138 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">Minerva Math</td>
|
| 139 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">4-shot</td>
|
| 140 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">9.96</td>
|
| 141 |
+
<td style="text-align:right; background-color: #DAE8FF; color: #2D2D2D;">11.5</td>
|
| 142 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">30.3</td>
|
| 143 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">21.3</td>
|
| 144 |
+
</tr>
|
| 145 |
+
<tr>
|
| 146 |
+
<td colspan="6" style="text-align:center; background-color: #FFFFFF; color: #2D2D2D; font-style:italic;">
|
| 147 |
+
Code Tasks
|
| 148 |
+
</td>
|
| 149 |
+
</tr>
|
| 150 |
+
<tr>
|
| 151 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">HumanEval</td>
|
| 152 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">pass@1 [StarCoder Prompt]</td>
|
| 153 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">34.6</td>
|
| 154 |
+
<td style="text-align:right; background-color: #DAE8FF; color: #2D2D2D;">35.61</td>
|
| 155 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">68.08</td>
|
| 156 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">68.26</td>
|
| 157 |
+
</tr>
|
| 158 |
+
<tr>
|
| 159 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">HumanEval</td>
|
| 160 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">pass@1</td>
|
| 161 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">32</td>
|
| 162 |
+
<td style="text-align:right; background-color: #DAE8FF; color: #2D2D2D;">34</td>
|
| 163 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">60</td>
|
| 164 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">59</td>
|
| 165 |
+
</tr>
|
| 166 |
+
<tr>
|
| 167 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">HumanEval+</td>
|
| 168 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">pass@1</td>
|
| 169 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">29</td>
|
| 170 |
+
<td style="text-align:right; background-color: #DAE8FF; color: #2D2D2D;">29</td>
|
| 171 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">57</td>
|
| 172 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">56</td>
|
| 173 |
+
</tr>
|
| 174 |
+
<tr>
|
| 175 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">MBPP</td>
|
| 176 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">pass@1</td>
|
| 177 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">45</td>
|
| 178 |
+
<td style="text-align:right; background-color: #DAE8FF; color: #2D2D2D;">17</td>
|
| 179 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">72</td>
|
| 180 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">65</td>
|
| 181 |
+
</tr>
|
| 182 |
+
<tr>
|
| 183 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">MBPP+</td>
|
| 184 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">pass@1</td>
|
| 185 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">38</td>
|
| 186 |
+
<td style="text-align:right; background-color: #DAE8FF; color: #2D2D2D;">16</td>
|
| 187 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">60</td>
|
| 188 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">54</td>
|
| 189 |
+
</tr>
|
| 190 |
+
<tr>
|
| 191 |
+
<td colspan="10" style="text-align:center; background-color: #FFFFFF; color: #2D2D2D; font-style:italic;">
|
| 192 |
+
Multilingual Tasks
|
| 193 |
+
</td>
|
| 194 |
+
</tr>
|
| 195 |
+
<tr>
|
| 196 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">MMMLU</td>
|
| 197 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">5-shot</td>
|
| 198 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">30.93</td>
|
| 199 |
+
<td style="text-align:right; background-color: #DAE8FF; color: #2D2D2D;">31.02</td>
|
| 200 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">46.73</td>
|
| 201 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">48.55</td>
|
| 202 |
+
</tr>
|
| 203 |
+
<tr>
|
| 204 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">INCLUDE</td>
|
| 205 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">5-shot</td>
|
| 206 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">27.32</td>
|
| 207 |
+
<td style="text-align:right; background-color: #DAE8FF; color: #2D2D2D;">29.26</td>
|
| 208 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">42.6</td>
|
| 209 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">43.8</td>
|
| 210 |
+
</tr>
|
| 211 |
+
<tr>
|
| 212 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">MGSM</td>
|
| 213 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">8-shot</td>
|
| 214 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">13.92</td>
|
| 215 |
+
<td style="text-align:right; background-color: #DAE8FF; color: #2D2D2D;">15.12</td>
|
| 216 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">46.96</td>
|
| 217 |
+
<td style="text-align:right; background-color: #FFFFFF; color: #2D2D2D;">41.52</td>
|
| 218 |
+
</tr>
|
| 219 |
+
</tbody></table>
|
| 220 |
+
|
| 221 |
+
<table>
|
| 222 |
+
<caption><b>Multilingual Benchmarks and thr included languages:</b></caption>
|
| 223 |
+
<thead>
|
| 224 |
+
<tr>
|
| 225 |
+
<th style="text-align:left; background-color: #001d6c; color: white;">Benchmarks</th>
|
| 226 |
+
<th style="text-align:left; background-color: #001d6c; color: white;"># Langs</th>
|
| 227 |
+
<th style="text-align:center; background-color: #001d6c; color: white;">Languages</th>
|
| 228 |
+
</tr>
|
| 229 |
+
</thead>
|
| 230 |
+
<tbody>
|
| 231 |
+
<tr>
|
| 232 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">MMMLU</td>
|
| 233 |
+
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">11</td>
|
| 234 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">ar, de, en, es, fr, ja, ko, pt, zh, bn, hi</td>
|
| 235 |
+
</tr>
|
| 236 |
+
<tr>
|
| 237 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">INCLUDE</td>
|
| 238 |
+
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">14</td>
|
| 239 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">hi, bn, ta, te, ar, de, es, fr, it, ja, ko, nl, pt, zh</td>
|
| 240 |
+
</tr>
|
| 241 |
+
<tr>
|
| 242 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">MGSM</td>
|
| 243 |
+
<td style="text-align:center; background-color: #FFFFFF; color: #2D2D2D;">5</td>
|
| 244 |
+
<td style="text-align:left; background-color: #FFFFFF; color: #2D2D2D;">en, es, fr, ja, zh</td>
|
| 245 |
+
</tr>
|
| 246 |
+
</tbody>
|
| 247 |
+
</table>
|
| 248 |
+
|
| 249 |
+
**Model Architecture:**
|
| 250 |
+
Granite-4.0-H-300M-Base is based on a decoder-only dense transformer architecture. Core components of this architecture are: GQA, Mamba2, MLP with SwiGLU, RMSNorm, and shared input/output embeddings.
|
| 251 |
+
|
| 252 |
+
<table>
|
| 253 |
+
<thead>
|
| 254 |
+
<tr>
|
| 255 |
+
<th style="text-align:left; background-color: #001d6c; color: white;">Model</th>
|
| 256 |
+
<th style="text-align:center; background-color: #001d6c; color: white;">300M Dense</th>
|
| 257 |
+
<th style="text-align:center; background-color: #001d6c; color: white;">H 300M Dense</th>
|
| 258 |
+
<th style="text-align:center; background-color: #001d6c; color: white;">1B Dense</th>
|
| 259 |
+
<th style="text-align:center; background-color: #001d6c; color: white;">H 1B Dense</th>
|
| 260 |
+
</tr></thead>
|
| 261 |
+
<tbody>
|
| 262 |
+
<tr>
|
| 263 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">Embedding size</td>
|
| 264 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">1024</td>
|
| 265 |
+
<td style="text-align:center; background-color: #DAE8FF; color: black;">768</td>
|
| 266 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">2048</td>
|
| 267 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">1536</td>
|
| 268 |
+
</tr>
|
| 269 |
+
<tr>
|
| 270 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">Number of layers</td>
|
| 271 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">28 attention</td>
|
| 272 |
+
<td style="text-align:center; background-color: #DAE8FF; color: black;">4 attention / 28 Mamba2</td>
|
| 273 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">40 attention</td>
|
| 274 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">4 attention / 36 Mamba2</td>
|
| 275 |
+
</tr>
|
| 276 |
+
<tr>
|
| 277 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">Attention head size</td>
|
| 278 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">64</td>
|
| 279 |
+
<td style="text-align:center; background-color: #DAE8FF; color: black;">64</td>
|
| 280 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">128</td>
|
| 281 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">128</td>
|
| 282 |
+
</tr>
|
| 283 |
+
<tr>
|
| 284 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">Number of attention heads</td>
|
| 285 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">16</td>
|
| 286 |
+
<td style="text-align:center; background-color: #DAE8FF; color: black;">12</td>
|
| 287 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">16</td>
|
| 288 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">12</td>
|
| 289 |
+
</tr>
|
| 290 |
+
<tr>
|
| 291 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">Number of KV heads</td>
|
| 292 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">4</td>
|
| 293 |
+
<td style="text-align:center; background-color: #DAE8FF; color: black;">4</td>
|
| 294 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">4</td>
|
| 295 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">4</td>
|
| 296 |
+
</tr>
|
| 297 |
+
<tr>
|
| 298 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">Mamba2 state size</td>
|
| 299 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">-</td>
|
| 300 |
+
<td style="text-align:center; background-color: #DAE8FF; color: black;">128</td>
|
| 301 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">-</td>
|
| 302 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">128</td>
|
| 303 |
+
</tr>
|
| 304 |
+
<tr>
|
| 305 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">Number of Mamba2 heads</td>
|
| 306 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">-</td>
|
| 307 |
+
<td style="text-align:center; background-color: #DAE8FF; color: black;">48</td>
|
| 308 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">-</td>
|
| 309 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">48</td>
|
| 310 |
+
</tr>
|
| 311 |
+
|
| 312 |
+
<tr>
|
| 313 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">MLP / Shared expert hidden size</td>
|
| 314 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">2048</td>
|
| 315 |
+
<td style="text-align:center; background-color: #DAE8FF; color: black;">2048</td>
|
| 316 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">4096</td>
|
| 317 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">4096</td>
|
| 318 |
+
</tr>
|
| 319 |
+
<tr>
|
| 320 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">Num. Experts</td>
|
| 321 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">-</td>
|
| 322 |
+
<td style="text-align:center; background-color: #DAE8FF; color: black;">-</td>
|
| 323 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">-</td>
|
| 324 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">-</td>
|
| 325 |
+
</tr>
|
| 326 |
+
<tr>
|
| 327 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">Num. active Experts</td>
|
| 328 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">-</td>
|
| 329 |
+
<td style="text-align:center; background-color: #DAE8FF; color: black;">-</td>
|
| 330 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">-</td>
|
| 331 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">-</td>
|
| 332 |
+
</tr>
|
| 333 |
+
<tr>
|
| 334 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">Expert hidden size</td>
|
| 335 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">-</td>
|
| 336 |
+
<td style="text-align:center; background-color: #DAE8FF; color: black;">-</td>
|
| 337 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">-</td>
|
| 338 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">-</td>
|
| 339 |
+
</tr>
|
| 340 |
+
<tr>
|
| 341 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">MLP activation</td>
|
| 342 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">SwiGLU</td>
|
| 343 |
+
<td style="text-align:center; background-color: #DAE8FF; color: black;">SwiGLU</td>
|
| 344 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">SwiGLU</td>
|
| 345 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">SwiGLU</td>
|
| 346 |
+
</tr>
|
| 347 |
+
|
| 348 |
+
<tr>
|
| 349 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">Sequence length</td>
|
| 350 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">32K</td>
|
| 351 |
+
<td style="text-align:center; background-color: #DAE8FF; color: black;">32K</td>
|
| 352 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">128K</td>
|
| 353 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">128K</td>
|
| 354 |
+
</tr>
|
| 355 |
+
<tr>
|
| 356 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">Position embedding</td>
|
| 357 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">RoPE</td>
|
| 358 |
+
<td style="text-align:center; background-color: #DAE8FF; color: black;">NoPE</td>
|
| 359 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">RoPE</td>
|
| 360 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">NoPE</td>
|
| 361 |
+
</tr>
|
| 362 |
+
<tr>
|
| 363 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;"># Parameters</td>
|
| 364 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">350M</td>
|
| 365 |
+
<td style="text-align:center; background-color: #DAE8FF; color: black;">340M</td>
|
| 366 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">1.6B</td>
|
| 367 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">1.5B</td>
|
| 368 |
+
</tr>
|
| 369 |
+
<tr>
|
| 370 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;"># Active parameters</td>
|
| 371 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">350M</td>
|
| 372 |
+
<td style="text-align:center; background-color: #DAE8FF; color: black;">340M</td>
|
| 373 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">1.6B</td>
|
| 374 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">1.5B</td>
|
| 375 |
+
</tr>
|
| 376 |
+
</tbody></table>
|
| 377 |
+
|
| 378 |
+
|
| 379 |
+
**Training Data:** This model is trained on a mix of open source and proprietary data following a four-stage training strategy.
|
| 380 |
+
|
| 381 |
+
<table>
|
| 382 |
+
<thead>
|
| 383 |
+
<tr>
|
| 384 |
+
<th style="text-align:left; background-color: #001d6c; color: white;">Stage</th>
|
| 385 |
+
<th style="text-align:left; background-color: #001d6c; color: white;">Characteristics</th>
|
| 386 |
+
<th style="text-align:center; background-color: #001d6c; color: white;">300M Dense</th>
|
| 387 |
+
<th style="text-align:center; background-color: #001d6c; color: white;">H 300M Dense</th>
|
| 388 |
+
<th style="text-align:center; background-color: #001d6c; color: white;">1B Dense</th>
|
| 389 |
+
<th style="text-align:center; background-color: #001d6c; color: white;">H 1B Dense</th>
|
| 390 |
+
</tr></thead>
|
| 391 |
+
<tbody>
|
| 392 |
+
<tr>
|
| 393 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">I</td>
|
| 394 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">General mixture of training data, warmup, and power scheduler for learning rate.</td>
|
| 395 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">10</td>
|
| 396 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">10</td>
|
| 397 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">10</td>
|
| 398 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">10</td>
|
| 399 |
+
</tr>
|
| 400 |
+
<tr>
|
| 401 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">II</td>
|
| 402 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">General mixture of training data with higher percentages of code and math with power scheduler for learning rate.</td>
|
| 403 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">2</td>
|
| 404 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">2</td>
|
| 405 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">2</td>
|
| 406 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">2</td>
|
| 407 |
+
</tr>
|
| 408 |
+
<tr>
|
| 409 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">III</td>
|
| 410 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">High quality training data, exponential decay of learning rate.</td>
|
| 411 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">2</td>
|
| 412 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">2</td>
|
| 413 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">2</td>
|
| 414 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">2</td>
|
| 415 |
+
</tr>
|
| 416 |
+
<tr>
|
| 417 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">IV</td>
|
| 418 |
+
<td style="text-align:left; background-color: #FFFFFF; color: black;">High quality training data, linear decay to zero for learning rate.</td>
|
| 419 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">0.5</td>
|
| 420 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">0.5</td>
|
| 421 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">0.5</td>
|
| 422 |
+
<td style="text-align:center; background-color: #FFFFFF; color: black;">0.5</td>
|
| 423 |
+
</tr>
|
| 424 |
+
</tbody></table>
|
| 425 |
+
|
| 426 |
+
**Infrastructure:**
|
| 427 |
+
We trained the Granite 4.0 Nano Language Models utilizing an NVIDIA GB200 NVL72 cluster hosted in CoreWeave. Intra-rack communication occurs via the 72-GPU NVLink domain, and a non-blocking, full Fat-Tree NDR 400 Gb/s InfiniBand network provides inter-rack communication. This cluster provides a scalable and efficient infrastructure for training our models over thousands of GPUs.
|
| 428 |
+
|
| 429 |
+
**Ethical Considerations and Limitations:**
|
| 430 |
+
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-4.0-H-300M-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-4.0-H-300M-Base model with ethical intentions and in a responsible way.
|
| 431 |
+
|
| 432 |
+
**Resources**
|
| 433 |
+
- ⭐️ Learn about the latest updates with Granite: https://www.ibm.com/granite
|
| 434 |
+
- 📄 Get started with tutorials, best practices, and prompt engineering advice: https://www.ibm.com/granite/docs/
|
| 435 |
+
- 💡 Learn about the latest Granite learning resources: https://github.com/ibm-granite-community/
|
model.sig
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"mediaType":"application/vnd.dev.sigstore.bundle.v0.3+json","verificationMaterial":{"certificate":{"rawBytes":"
|
|
|
|
| 1 |
+
{"mediaType":"application/vnd.dev.sigstore.bundle.v0.3+json","verificationMaterial":{"certificate":{"rawBytes":"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"},"tlogEntries":[{"logIndex":"622898057","logId":{"keyId":"wNI9atQGlz+VWfO6LRygH4QUfY/8W4RFwiT5i5WRgB0="},"kindVersion":{"kind":"dsse","version":"0.0.1"},"integratedTime":"1760958400","inclusionPromise":{"signedEntryTimestamp":"MEQCIC/onAtZ9JDPNAPX3kcSr0EAifZD/Usp+i+VkRnS199kAiA1oFi27T82g7mmLjIABQaCtRqu5KnjllMOWwoqMs4WIw=="},"inclusionProof":{"logIndex":"500993795","rootHash":"vFVYmsmhzvStgPMVDyJ9pvfw5j+LIy99hCfJ6CzA35k=","treeSize":"500993797","hashes":["J1w3BS+nwpfoA2A12MuNx41yW/rUWIhfrw6Hpu62lVA=","f8x4TQBEJk9ITPwdNxllZjilTO1WAHYyxSTkAQAYlxw=","HRunOYKRvlBcrL8ZB7b08zZaSZs+ep77ffmowOQP6ss=","g7fw06H7gWCS4czPCtDywPGjynIKU9L8icKBJpnftA0=","E6wBfBtRibnjOOTpsw7AglI6qHKKYQLbBsUkCOHSOYg=","ReikE5f8Mt9Fxmn0HkbkyFShtaBCm7pDV+htSCsWxbo=","+IO+Gs7LjN0FE0XQsQ5DJ8OGQCAzMCgmYN7++VB2AxA=","sZz8mjhiYp9OEs0f1dQyfPRcCGwvwMhxb69tbtxYGhA=","IEfBUPr/rhNZzb8B9lMr45R1g8+4y8KbM+K61qt/yEY=","zz0u1VmttTdodFw2vwldBHWAinT4XoiUwP6qLYf3ldE=","ignBbwyLkAXSQfQoWjGMZ2m2p9d7dDqWaXW4XAlV30g=","Uh9lxATVI904AWJn3ydpTnpBfb8dfuOMpPBBb10+kv4=","vCYQZqPwuKPvCNeCqAVIBfGK0xps0RT3XAe3bs7RKyM=","eGvIh/2VuASWOIw2cAlggnhHIj5WniQ0TmehEpx+ZbU=","2Wv4GiithwNukRKV06clevnQQYCzXmSS/+/OJtXgsXQ=","1mfy94KpcItqshH9+gwqV6jccupcaMpVsF28New8zDY=","vS7O4ozHIQZJWBiov+mkpI27GE8zAmVCEkRcP3NDyNE="],"checkpoint":{"envelope":"rekor.sigstore.dev - 1193050959916656506\n500993797\nvFVYmsmhzvStgPMVDyJ9pvfw5j+LIy99hCfJ6CzA35k=\n\n— rekor.sigstore.dev wNI9ajBGAiEA7d8CKKVz8h7r9+pMWzTtxvS3GQ+MQdCvvzhu/DVCvgkCIQD+4UFlxjskVqBAM0JgJBEboVIeFImyK/Up2EV/11+pZw==\n"}},"canonicalizedBody":"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"}],"timestampVerificationData":{"rfc3161Timestamps":[{"signedTimestamp":"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"}]}},"dsseEnvelope":{"payload":"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","payloadType":"application/vnd.in-toto+json","signatures":[{"sig":"MEYCIQC8nMeVQpYW6PmTzzC8S4mkHW94KgvuUtBBXs3X7Lk6NgIhAPGYAxf/KS1O0JPLSVNMC3oOoHYtis3HWFq5RaQcV78D"}]}}
|