Text Generation
Transformers
Safetensors
mistral3
image-text-to-text
decision-model
typed-decisions
jev
jevbench
fp8
compressed-tensors
vision-language
conversational
Instructions to use StandardThinking/StandardOne-8B-FP8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use StandardThinking/StandardOne-8B-FP8 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="StandardThinking/StandardOne-8B-FP8") messages = [ { "role": "user", "content": [ {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"}, {"type": "text", "text": "What animal is on the candy?"} ] }, ] pipe(text=messages)# pip install -U transformers accelerate # Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("StandardThinking/StandardOne-8B-FP8") model = AutoModelForMultimodalLM.from_pretrained("StandardThinking/StandardOne-8B-FP8", device_map="auto") messages = [ { "role": "user", "content": [ {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"}, {"type": "text", "text": "What animal is on the candy?"} ] }, ] inputs = processor.apply_chat_template( messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt", ).to(model.device) outputs = model.generate(**inputs, max_new_tokens=256) print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use StandardThinking/StandardOne-8B-FP8 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "StandardThinking/StandardOne-8B-FP8" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "StandardThinking/StandardOne-8B-FP8", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/StandardThinking/StandardOne-8B-FP8
- SGLang
How to use StandardThinking/StandardOne-8B-FP8 with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "StandardThinking/StandardOne-8B-FP8" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "StandardThinking/StandardOne-8B-FP8", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "StandardThinking/StandardOne-8B-FP8" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "StandardThinking/StandardOne-8B-FP8", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use StandardThinking/StandardOne-8B-FP8 with Docker Model Runner:
docker model run hf.co/StandardThinking/StandardOne-8B-FP8
Update weights
Browse files- SHA256SUMS +3 -3
- model.safetensors +1 -1
- release-manifest.json +54 -22
SHA256SUMS
CHANGED
|
@@ -1,15 +1,15 @@
|
|
| 1 |
1495e1e757ef4d0925a2350563cf5754bb23c51701a8ec4fb3c5cdcbedae6747 LICENSE
|
| 2 |
20486cf6b2e42440b0aea913ea475c818664f34e76e4501651d6cb21987a508b NOTICE
|
| 3 |
-
|
| 4 |
66e0fda8c8ab269ccd010dca6196931d475a22c3f2eebe5b6b3a2ec33a354d0e SYSTEM_PROMPT.txt
|
| 5 |
74eeb55fd3341286ec3fd44e902b7120721acc81cd394e96b431f85e93a1ea56 chat_template.jinja
|
| 6 |
41c931acee9e6c9842e04920d17715d249c10ec6a3e47e8eeacabec609a8066d config.json
|
| 7 |
e0923390059f84a9180b00e5501778acc45ea9856cd7f2fd68208b360927c677 generation_config.json
|
| 8 |
-
|
| 9 |
81b8377f36c5b3d60900b333214d85160a0500b84576e969092c6fd214f69538 params.json
|
| 10 |
ece2373c2ae391bce18785a1810543bc6173a1f28b6767cffab2c35dbea5f002 processor_config.json
|
| 11 |
32fb95d7d1f6b64b205a327074072dc777b98acab8beaaa031806474bf406ccb recipe.yaml
|
| 12 |
-
|
| 13 |
0a5c981e8c5c6f8886ee007a6d4543a0be6b221cb9ca32a8709384a4c6fc8cbb special_tokens_map.json
|
| 14 |
600bb27946565481ecf51ba8aee252e49b9a68507866080ac9c30185bb312843 tekken.json
|
| 15 |
d5f6046775b112f0e2d456ee9dba450684ab964fe5c4e231599bdc6773028135 tokenizer.json
|
|
|
|
| 1 |
1495e1e757ef4d0925a2350563cf5754bb23c51701a8ec4fb3c5cdcbedae6747 LICENSE
|
| 2 |
20486cf6b2e42440b0aea913ea475c818664f34e76e4501651d6cb21987a508b NOTICE
|
| 3 |
+
6375bc87de41a64a22930e8244b44aa8d19397607035a616140421b56bd8220e README.md
|
| 4 |
66e0fda8c8ab269ccd010dca6196931d475a22c3f2eebe5b6b3a2ec33a354d0e SYSTEM_PROMPT.txt
|
| 5 |
74eeb55fd3341286ec3fd44e902b7120721acc81cd394e96b431f85e93a1ea56 chat_template.jinja
|
| 6 |
41c931acee9e6c9842e04920d17715d249c10ec6a3e47e8eeacabec609a8066d config.json
|
| 7 |
e0923390059f84a9180b00e5501778acc45ea9856cd7f2fd68208b360927c677 generation_config.json
|
| 8 |
+
78088aa5720b56d417ed980ab605a9919d36ede7852052c2c34483a813040bda model.safetensors
|
| 9 |
81b8377f36c5b3d60900b333214d85160a0500b84576e969092c6fd214f69538 params.json
|
| 10 |
ece2373c2ae391bce18785a1810543bc6173a1f28b6767cffab2c35dbea5f002 processor_config.json
|
| 11 |
32fb95d7d1f6b64b205a327074072dc777b98acab8beaaa031806474bf406ccb recipe.yaml
|
| 12 |
+
386cb0ca65a10dbc1490bf742decbff3dce2ccc0f06d7424bb95c15690ffe7a4 release-manifest.json
|
| 13 |
0a5c981e8c5c6f8886ee007a6d4543a0be6b221cb9ca32a8709384a4c6fc8cbb special_tokens_map.json
|
| 14 |
600bb27946565481ecf51ba8aee252e49b9a68507866080ac9c30185bb312843 tekken.json
|
| 15 |
d5f6046775b112f0e2d456ee9dba450684ab964fe5c4e231599bdc6773028135 tokenizer.json
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 10423550984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:78088aa5720b56d417ed980ab605a9919d36ede7852052c2c34483a813040bda
|
| 3 |
size 10423550984
|
release-manifest.json
CHANGED
|
@@ -2,7 +2,7 @@
|
|
| 2 |
"release": "StandardOne-8B-FP8",
|
| 3 |
"source": {
|
| 4 |
"checkpoint": "StandardOne-8B",
|
| 5 |
-
"revision": "
|
| 6 |
},
|
| 7 |
"quantization": {
|
| 8 |
"tool": "llmcompressor",
|
|
@@ -39,29 +39,61 @@
|
|
| 39 |
"transformers": "5.17.0",
|
| 40 |
"compressed_tensors": "0.19.0"
|
| 41 |
},
|
| 42 |
-
"date": "2026-
|
| 43 |
},
|
| 44 |
"validation": {
|
| 45 |
-
"path": "SGLang 0.5.20 + jev-adapter
|
|
|
|
|
|
|
| 46 |
"accuracy_pct": {
|
| 47 |
-
"
|
| 48 |
-
|
| 49 |
-
|
| 50 |
-
|
| 51 |
-
|
| 52 |
-
"
|
| 53 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 54 |
},
|
| 55 |
-
"
|
| 56 |
-
|
| 57 |
-
|
| 58 |
-
|
| 59 |
-
"judge-proxy-v1": 90.5,
|
| 60 |
-
"internal-eval-v2-realistic": 90.33,
|
| 61 |
-
"distprob-val-v2": 81.37,
|
| 62 |
-
"internal-eval-hardproxy-v1": 52.5
|
| 63 |
-
},
|
| 64 |
-
"fitted_temperature": 1.4,
|
| 65 |
-
"bf16_release_temperature": 1.65
|
| 66 |
-
}
|
| 67 |
}
|
|
|
|
| 2 |
"release": "StandardOne-8B-FP8",
|
| 3 |
"source": {
|
| 4 |
"checkpoint": "StandardOne-8B",
|
| 5 |
+
"revision": "2b31eebec5ce6aa4c023c8e19f1af9b2e2688200"
|
| 6 |
},
|
| 7 |
"quantization": {
|
| 8 |
"tool": "llmcompressor",
|
|
|
|
| 39 |
"transformers": "5.17.0",
|
| 40 |
"compressed_tensors": "0.19.0"
|
| 41 |
},
|
| 42 |
+
"date": "2026-10-04"
|
| 43 |
},
|
| 44 |
"validation": {
|
| 45 |
+
"path": "SGLang 0.5.20 + jev-adapter (label scheme upper), served wording, single option order, default temperature 1.65; accuracy is argmax",
|
| 46 |
+
"reference": "StandardOne-8B v2.2 BF16 through the same path",
|
| 47 |
+
"overall_argmax_agreement_pct": 95.83,
|
| 48 |
"accuracy_pct": {
|
| 49 |
+
"many-option questions, 53–151 options": {
|
| 50 |
+
"n": 18000,
|
| 51 |
+
"bf16": 82.67,
|
| 52 |
+
"fp8": 82.34
|
| 53 |
+
},
|
| 54 |
+
"the same question set, at most 26 options": {
|
| 55 |
+
"n": 750,
|
| 56 |
+
"bf16": 88.13,
|
| 57 |
+
"fp8": 88.13
|
| 58 |
+
},
|
| 59 |
+
"long-document questions": {
|
| 60 |
+
"n": 150,
|
| 61 |
+
"bf16": 40.0,
|
| 62 |
+
"fp8": 40.67
|
| 63 |
+
},
|
| 64 |
+
"held-out decision set": {
|
| 65 |
+
"n": 600,
|
| 66 |
+
"bf16": 82.5,
|
| 67 |
+
"fp8": 84.0
|
| 68 |
+
},
|
| 69 |
+
"hard proxy": {
|
| 70 |
+
"n": 600,
|
| 71 |
+
"bf16": 53.5,
|
| 72 |
+
"fp8": 52.67
|
| 73 |
+
},
|
| 74 |
+
"realistic transfer set": {
|
| 75 |
+
"n": 600,
|
| 76 |
+
"bf16": 91.5,
|
| 77 |
+
"fp8": 90.83
|
| 78 |
+
},
|
| 79 |
+
"JevBench public easy": {
|
| 80 |
+
"n": 48,
|
| 81 |
+
"bf16": 100.0,
|
| 82 |
+
"fp8": 100.0
|
| 83 |
+
},
|
| 84 |
+
"JevBench public standard": {
|
| 85 |
+
"n": 72,
|
| 86 |
+
"bf16": 98.61,
|
| 87 |
+
"fp8": 97.22
|
| 88 |
+
},
|
| 89 |
+
"JevBench public hard": {
|
| 90 |
+
"n": 111,
|
| 91 |
+
"bf16": 56.76,
|
| 92 |
+
"fp8": 55.86
|
| 93 |
+
}
|
| 94 |
},
|
| 95 |
+
"questions_compared_for_agreement": 28406,
|
| 96 |
+
"temperature": "not refit for this build; use the StandardOne-8B serving settings"
|
| 97 |
+
},
|
| 98 |
+
"version": "v2.2"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 99 |
}
|