File size: 15,576 Bytes
5fdd752 0b2da09 5fdd752 0b2da09 4981e98 0b2da09 4981e98 0b2da09 41de52e 34f204f 0b2da09 ce48396 0b2da09 ce48396 0b2da09 5f1d00a 0b2da09 602396c 0b2da09 ce48396 0b2da09 ce48396 0b2da09 ce48396 0b2da09 602396c 0b2da09 4981e98 0b2da09 4981e98 0b2da09 41de52e 4981e98 0b2da09 ce48396 0b2da09 ce48396 0b2da09 ce48396 0b2da09 ce48396 0b2da09 2f2a1b4 3db7e6f 5fdd752 9614a0b 11b59d3 3db7e6f 9614a0b |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 |
---
license: apache-2.0
pipeline_tag: text-generation
tags:
- fp8
- quantized
- llm-compressor
- compressed-tensors
- red hat
base_model:
- ibm-granite/granite-4.0-h-small
---
# Granite-4.0-h-small
## Model Overview
- **Model Architecture:** GraniteMoeHybridForCausalLM
- **Input:** Text
- **Output:** Text
- **Model Optimizations:**
- **Weight quantization:** FP8
- **Activation quantization:** FP8
- **Release Date:**
- **Version:** 1.0
- **Model Developers:**: Red Hat
Quantized version of [ibm-granite/granite-4.0-h-small](https://huggingface.co/ibm-granite/granite-4.0-h-small).
### Model Optimizations
This model was obtained by quantizing the weights and activations of [ibm-granite/granite-4.0-h-small](https://huggingface.co/ibm-granite/granite-4.0-h-small) to FP8 data type.
This optimization reduces the number of bits per parameter from 16 to 8, reducing the disk size and GPU memory requirements by approximately 50%.
Only the weights and activations of the linear operators within transformers blocks of the language model are quantized.
## Deployment
### Use with vLLM
1. Install vLLM from main:
```
uv pip install -U git+https://github.com/vllm-project/vllm.git \
--extra-index-url https://wheels.vllm.ai/nightly \
--no-deps \
--no-cache
uv pip install compressed-tensors==0.12.3a20251114 --no-cache
uv pip install --upgrade torchvision --break-system-packages --no-cache
uv pip install cloudpickle msgspec zmq blake3 cachetools prometheus_client fastapi openai openai_harmony pybase64 llguidance diskcache xgrammar lm-format-enforcer partial-json-parser cbor2 einops gguf numba --no-cache
```
2. Initialize vLLM server:
```
vllm serve RedHatAI/granite-4.0-h-small-FP8-dynamic --tensor_parallel_size 1
```
3. Send requests to the server:
```python
from openai import OpenAI
# Modify OpenAI's API key and API base to use vLLM's API server.
openai_api_key = "EMPTY"
openai_api_base = "http://<your-server-host>:8000/v1"
client = OpenAI(
api_key=openai_api_key,
base_url=openai_api_base,
)
model = "RedHatAI/granite-4.0-h-small-FP8-dynamic"
messages = [
{"role": "user", "content": "Explain quantum mechanics clearly and concisely."},
]
outputs = client.chat.completions.create(
model=model,
messages=messages,
)
generated_text = outputs.choices[0].message.content
print(generated_text)
```
## Creation
This model was quantized using the [llm-compressor](https://github.com/vllm-project/llm-compressor) library as shown below.
<details>
<summary>Creation details</summary>
Install specific version:
```
uv pip install git+https://github.com/vllm-project/llm-compressor.git@refs/pull/2001/head --no-cache
uv pip install --upgrade torchvision --break-system-packages --no-cache
```
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
from llmcompressor import oneshot
from llmcompressor.modifiers.quantization import QuantizationModifier
from llmcompressor.utils import dispatch_for_generation
from llmcompressor.modeling import replace_modules_for_calibration
from llmcompressor.modeling.granite4 import pack_3d_experts
MODEL_ID = "ibm-granite/granite-4.0-h-small"
model = AutoModelForCausalLM.from_pretrained(MODEL_ID, torch_dtype="auto")
tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
model = replace_modules_for_calibration(model)
ignore_lay = ["lm_head", "re:.*block_sparse_moe.router"]
recipe = QuantizationModifier(
targets=["Linear"],
scheme="FP8_DYNAMIC",
ignore=ignore_lay,
)
oneshot(model=model, recipe=recipe)
print("========== SAMPLE GENERATION ==============")
dispatch_for_generation(model)
input_ids = tokenizer(
"Describe Large Language Model", return_tensors="pt"
).input_ids.to(model.device)
output = model.generate(input_ids, max_new_tokens=35)
print(tokenizer.decode(output[0]))
print("==========================================")
SAVE_DIR = MODEL_ID.rstrip("/").split("/")[-1] + "-FP8-dynamic"
print(f"Saving to {SAVE_DIR}")
model.save_pretrained(SAVE_DIR)
tokenizer.save_pretrained(SAVE_DIR)
pack_3d_experts(SAVE_DIR)
```
</details>
## Evaluation
The model was evaluated on the OpenLLM leaderboard task, using [lm-evaluation-harness](https://github.com/EleutherAI/lm-evaluation-harness).
[vLLM](https://docs.vllm.ai/en/stable/) was used for all evaluations.
<details>
<summary>Evaluation details</summary>
Install vLLM from main:
```
uv pip install -U git+https://github.com/vllm-project/vllm.git \
--extra-index-url https://wheels.vllm.ai/nightly \
--no-deps \
--no-cache
uv pip install compressed-tensors==0.12.3a20251114 --no-cache
uv pip install --upgrade torchvision --break-system-packages --no-cache
uv pip install cloudpickle msgspec zmq blake3 cachetools prometheus_client fastapi openai openai_harmony pybase64 llguidance diskcache xgrammar lm-format-enforcer partial-json-parser cbor2 einops gguf numba --no-cache
```
**Openllm V1**
```
lm_eval \
--model vllm \
--model_args pretrained="RedHatAI/granite-4.0-h-small-FP8-dynamic",dtype=auto,add_bos_token=True,max_model_len=16384,tensor_parallel_size=1,gpu_memory_utilization=0.9,enable_chunked_prefill=True,trust_remote_code=True \
--tasks openllm \
--write_out \
--batch_size auto \
--show_config
```
**Openllm V2**
```
lm_eval \
--model vllm \
--model_args pretrained="RedHatAI/granite-4.0-h-small-FP8-dynamic",dtype=auto,add_bos_token=False,max_model_len=16384,tensor_parallel_size=1,gpu_memory_utilization=0.7,disable_log_stats=True,enable_chunked_prefill=True,trust_remote_code=True \
--tasks leaderboard \
--apply_chat_template \
--fewshot_as_multiturn \
--write_out \
--batch_size auto \
--show_config
```
**Coding Benchmarks**
```
evalplus.evaluate --model "RedHatAI/granite-4.0-h-small-FP8-dynamic" \
--dataset "humaneval" \
--backend vllm \
--tp 1 \
--greedy
evalplus.evaluate --model "RedHatAI/granite-4.0-h-small-FP8-dynamic" \
--dataset "mbpp" \
--backend vllm \
--tp 1 \
--greedy
```
</details>
### Accuracy Comparison
<table>
<thead>
<tr>
<th>Category</th>
<th>Metric</th>
<th>ibm-granite/granite-4.0-h-small</th>
<th>ibm-granite/granite-4.0-h-small-FP8</th>
<th>RedHatAI/granite-4.0-h-small-FP8-block</th>
<th>RedHatAI/granite-4.0-h-small-FP8-dynamic</th>
</tr>
</thead>
<tbody>
<!-- OpenLLM Leaderboard V1 -->
<tr>
<td rowspan="7"><b>OpenLLM V1</b></td>
<td>ARC-Challenge (Acc-Norm, 25-shot)</td>
<td>72.27</td>
<td>72.10 (99.76%)</td>
<td>72.27 (100.00%)</td>
<td>72.10 (99.76%)</td>
</tr>
<tr>
<td>GSM8K (Strict-Match, 5-shot)</td>
<td>85.22</td>
<td>85.29 (100.09%)</td>
<td>85.52 (100.36%)</td>
<td>84.84 (99.56%)</td>
</tr>
<tr>
<td>HellaSwag (Acc-Norm, 10-shot)</td>
<td>86.08</td>
<td>85.88 (99.77%)</td>
<td>85.96 (99.86%)</td>
<td>85.88 (99.77%)</td>
</tr>
<tr>
<td>MMLU (Acc, 5-shot)</td>
<td>77.15</td>
<td>77.18 (100.03%)</td>
<td>77.23 (100.09%)</td>
<td>77.18 (100.03%)</td>
</tr>
<tr>
<td>TruthfulQA (MC2, 0-shot)</td>
<td>57.64</td>
<td>57.63 (99.99%)</td>
<td>57.94 (100.52%)</td>
<td>57.63 (100.00%)</td>
</tr>
<tr>
<td>Winogrande (Acc, 5-shot)</td>
<td>81.37</td>
<td>81.45 (100.10%)</td>
<td>80.82 (99.32%)</td>
<td>81.45 (100.10%)</td>
</tr>
<tr>
<td><b>Average Score</b></td>
<td><b>76.62</b></td>
<td><b>76.59 (99.96%)</b></td>
<td><b>76.62 (100.00%)</b></td>
<td><b>76.51 (99.86%)</b></td>
</tr>
<!-- OpenLLM Leaderboard V2 -->
<tr>
<td rowspan="7"><b>OpenLLM V2</b></td>
<td>IFEval (Inst Level Strict Acc, 0-shot)</td>
<td>87.53</td>
<td>87.17 (99.59%)</td>
<td>86.69 (99.04%)</td>
<td>87.41 (99.86%)</td>
</tr>
<tr>
<td>BBH (Acc-Norm, 3-shot)</td>
<td>61.52</td>
<td>61.31 (99.66%)</td>
<td>61.40 (99.80%)</td>
<td>61.19 (99.46%)</td>
</tr>
<tr>
<td>Math-Hard (Exact-Match, 4-shot)</td>
<td>46.22</td>
<td>43.73 (94.61%)</td>
<td>43.88 (94.93%)</td>
<td>41.77 (90.36%)</td>
</tr>
<tr>
<td>GPQA (Acc-Norm, 0-shot)</td>
<td>35.23</td>
<td>34.98 (99.29%)</td>
<td>34.23 (97.14%)</td>
<td>34.23 (97.14%)</td>
</tr>
<tr>
<td>MUSR (Acc-Norm, 0-shot)</td>
<td>46.69</td>
<td>46.56 (99.72%)</td>
<td>45.77 (98.02%)</td>
<td>45.77 (98.02%)</td>
</tr>
<tr>
<td>MMLU-Pro (Acc, 5-shot)</td>
<td>47.99</td>
<td>47.63 (99.26%)</td>
<td>47.93 (99.88%)</td>
<td>47.58 (99.15%)</td>
</tr>
<tr>
<td><b>Average Score</b></td>
<td><b>54.20</b></td>
<td><b>53.56 (98.82%)</b></td>
<td><b>53.32 (98.38%)</b></td>
<td><b>52.99 (97.77%)</b></td>
</tr>
</tbody>
</table>
<!--
### Accuracy
<table>
<thead>
<tr>
<th>Category</th>
<th>Metric</th>
<th>ibm-granite/granite-4.0-h-small</th>
<th>ibm-granite/granite-4.0-h-small-FP8</th>
<th>Recovery (%)</th>
</tr>
</thead>
<tbody>
<tr>
<td rowspan="7"><b>OpenLLM V1</b></td>
<td>ARC-Challenge (Acc-Norm, 25-shot)</td>
<td>72.27</td>
<td>72.10</td>
<td>99.76</td>
</tr>
<tr>
<td>GSM8K (Strict-Match, 5-shot)</td>
<td>85.22</td>
<td>85.29</td>
<td>100.09</td>
</tr>
<tr>
<td>HellaSwag (Acc-Norm, 10-shot)</td>
<td>86.08</td>
<td>85.88</td>
<td>99.77</td>
</tr>
<tr>
<td>MMLU (Acc, 5-shot)</td>
<td>77.15</td>
<td>77.18</td>
<td>100.03</td>
</tr>
<tr>
<td>TruthfulQA (MC2, 0-shot)</td>
<td>57.64</td>
<td>57.63</td>
<td>99.99</td>
</tr>
<tr>
<td>Winogrande (Acc, 5-shot)</td>
<td>81.37</td>
<td>81.45</td>
<td>100.10</td>
</tr>
<tr>
<td><b>Average Score</b></td>
<td><b>76.62</b></td>
<td><b>76.59</b></td>
<td><b>99.96</b></td>
</tr>
<tr>
<td rowspan="7"><b>OpenLLM V2</b></td>
<td>IFEval (Inst Level Strict Acc, 0-shot)</td>
<td>87.53</td>
<td>87.17</td>
<td>99.59</td>
</tr>
<tr>
<td>BBH (Acc-Norm, 3-shot)</td>
<td>61.52</td>
<td>61.31</td>
<td>99.66</td>
</tr>
<tr>
<td>Math-Hard (Exact-Match, 4-shot)</td>
<td>46.22</td>
<td>43.73</td>
<td>94.61</td>
</tr>
<tr>
<td>GPQA (Acc-Norm, 0-shot)</td>
<td>35.23</td>
<td>34.98</td>
<td>99.29</td>
</tr>
<tr>
<td>MUSR (Acc-Norm, 0-shot)</td>
<td>46.69</td>
<td>46.56</td>
<td>99.72</td>
</tr>
<tr>
<td>MMLU-Pro (Acc, 5-shot)</td>
<td>47.99</td>
<td>47.63</td>
<td>99.26</td>
</tr>
<tr>
<td><b>Average Score</b></td>
<td><b>54.20</b></td>
<td><b>53.56</b></td>
<td><b>98.82</b></td>
</tr>
</tbody>
</table>
### Accuracy
<table>
<thead>
<tr>
<th>Category</th>
<th>Metric</th>
<th>ibm-granite/granite-4.0-h-small</th>
<th>RedHatAI/granite-4.0-h-small-FP8-block</th>
<th>Recovery (%)</th>
</tr>
</thead>
<tbody>
<tr>
<td rowspan="7"><b>OpenLLM V1</b></td>
<td>ARC-Challenge (Acc-Norm, 25-shot)</td>
<td>72.27</td>
<td>72.27</td>
<td>100.00</td>
</tr>
<tr>
<td>GSM8K (Strict-Match, 5-shot)</td>
<td>85.22</td>
<td>85.52</td>
<td>100.36</td>
</tr>
<tr>
<td>HellaSwag (Acc-Norm, 10-shot)</td>
<td>86.08</td>
<td>85.96</td>
<td>99.86</td>
</tr>
<tr>
<td>MMLU (Acc, 5-shot)</td>
<td>77.15</td>
<td>77.23</td>
<td>100.09</td>
</tr>
<tr>
<td>TruthfulQA (MC2, 0-shot)</td>
<td>57.64</td>
<td>57.94</td>
<td>100.52</td>
</tr>
<tr>
<td>Winogrande (Acc, 5-shot)</td>
<td>81.37</td>
<td>80.82</td>
<td>99.32</td>
</tr>
<tr>
<td><b>Average Score</b></td>
<td><b>76.62</b></td>
<td><b>76.62</b></td>
<td><b>100.00</b></td>
</tr>
<tr>
<td rowspan="7"><b>OpenLLM V2</b></td>
<td>IFEval (Inst Level Strict Acc, 0-shot)</td>
<td>87.53</td>
<td>86.69</td>
<td>99.04</td>
</tr>
<tr>
<td>BBH (Acc-Norm, 3-shot)</td>
<td>61.52</td>
<td>61.40</td>
<td>99.80</td>
</tr>
<tr>
<td>Math-Hard (Exact-Match, 4-shot)</td>
<td>46.22</td>
<td>43.88</td>
<td>94.93</td>
</tr>
<tr>
<td>GPQA (Acc-Norm, 0-shot)</td>
<td>35.23</td>
<td>34.23</td>
<td>97.14</td>
</tr>
<tr>
<td>MUSR (Acc-Norm, 0-shot)</td>
<td>46.69</td>
<td>45.77</td>
<td>98.02</td>
</tr>
<tr>
<td>MMLU-Pro (Acc, 5-shot)</td>
<td>47.99</td>
<td>47.93</td>
<td>99.88</td>
</tr>
<tr>
<td><b>Average Score</b></td>
<td><b>54.20</b></td>
<td><b>53.32</b></td>
<td><b>98.38</b></td>
</tr>
</tbody>
</table>
### Accuracy
<table>
<thead>
<tr>
<th>Category</th>
<th>Metric</th>
<th>ibm-granite/granite-4.0-h-small</th>
<th>RedHatAI/granite-4.0-h-small-FP8-dynamic</th>
<th>Recovery (%)</th>
</tr>
</thead>
<tbody>
<tr>
<td rowspan="7"><b>OpenLLM V1</b></td>
<td>ARC-Challenge (Acc-Norm, 25-shot)</td>
<td>72.27</td>
<td>72.10</td>
<td>99.76</td>
</tr>
<tr>
<td>GSM8K (Strict-Match, 5-shot)</td>
<td>85.22</td>
<td>84.84</td>
<td>99.56</td>
</tr>
<tr>
<td>HellaSwag (Acc-Norm, 10-shot)</td>
<td>86.08</td>
<td>85.88</td>
<td>99.77</td>
</tr>
<tr>
<td>MMLU (Acc, 5-shot)</td>
<td>77.15</td>
<td>77.18</td>
<td>100.03</td>
</tr>
<tr>
<td>TruthfulQA (MC2, 0-shot)</td>
<td>57.64</td>
<td>57.63</td>
<td>100.00</td>
</tr>
<tr>
<td>Winogrande (Acc, 5-shot)</td>
<td>81.37</td>
<td>81.45</td>
<td>100.10</td>
</tr>
<tr>
<td><b>Average Score</b></td>
<td><b>76.62</b></td>
<td><b>76.51</b></td>
<td><b>99.86</b></td>
</tr>
<tr>
<td rowspan="7"><b>OpenLLM V2</b></td>
<td>IFEval (Inst Level Strict Acc, 0-shot)</td>
<td>87.53</td>
<td>87.41</td>
<td>99.86</td>
</tr>
<tr>
<td>BBH (Acc-Norm, 3-shot)</td>
<td>61.52</td>
<td>61.19</td>
<td>99.46</td>
</tr>
<tr>
<td>Math-Hard (Exact-Match, 4-shot)</td>
<td>46.22</td>
<td>41.77</td>
<td>90.36</td>
</tr>
<tr>
<td>GPQA (Acc-Norm, 0-shot)</td>
<td>35.23</td>
<td>34.23</td>
<td>97.14</td>
</tr>
<tr>
<td>MUSR (Acc-Norm, 0-shot)</td>
<td>46.69</td>
<td>45.77</td>
<td>98.02</td>
</tr>
<tr>
<td>MMLU-Pro (Acc, 5-shot)</td>
<td>47.99</td>
<td>47.58</td>
<td>99.15</td>
</tr>
<tr>
<td><b>Average Score</b></td>
<td><b>54.20</b></td>
<td><b>52.99</b></td>
<td><b>97.77</b></td>
</tr>
</tbody>
</table> -->
|