Instructions to use optimum-intel-internal-testing/tiny-random-gpt-oss-mxfp4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use optimum-intel-internal-testing/tiny-random-gpt-oss-mxfp4 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="optimum-intel-internal-testing/tiny-random-gpt-oss-mxfp4")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("optimum-intel-internal-testing/tiny-random-gpt-oss-mxfp4", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use optimum-intel-internal-testing/tiny-random-gpt-oss-mxfp4 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "optimum-intel-internal-testing/tiny-random-gpt-oss-mxfp4"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "optimum-intel-internal-testing/tiny-random-gpt-oss-mxfp4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/optimum-intel-internal-testing/tiny-random-gpt-oss-mxfp4

SGLang

How to use optimum-intel-internal-testing/tiny-random-gpt-oss-mxfp4 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 "optimum-intel-internal-testing/tiny-random-gpt-oss-mxfp4" \
    --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": "optimum-intel-internal-testing/tiny-random-gpt-oss-mxfp4",
		"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 "optimum-intel-internal-testing/tiny-random-gpt-oss-mxfp4" \
        --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": "optimum-intel-internal-testing/tiny-random-gpt-oss-mxfp4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use optimum-intel-internal-testing/tiny-random-gpt-oss-mxfp4 with Docker Model Runner:
```
docker model run hf.co/optimum-intel-internal-testing/tiny-random-gpt-oss-mxfp4
```

Fix: add rope_theta inside rope_scaling for transformers >= v5.4 compatibility

by echarlaix HF Staff - opened Jun 4

base: refs/heads/main

←

from: refs/pr/2

Discussion Files changed

-50

Files changed (1) hide show

config.json +3 -50

config.json CHANGED Viewed

@@ -1,50 +1,3 @@
-{
-  "architectures": [
-    "GptOssForCausalLM"
-  ],
-  "attention_bias": true,
-  "attention_dropout": 0.0,
-  "dtype": "bfloat16",
-  "eos_token_id": 200002,
-  "experts_per_token": 4,
-  "head_dim": 32,
-  "hidden_act": "silu",
-  "hidden_size": 32,
-  "initial_context_length": 4096,
-  "initializer_range": 0.02,
-  "intermediate_size": 64,
-  "layer_types": [
-    "sliding_attention",
-    "full_attention"
-  ],
-  "max_position_embeddings": 131072,
-  "model_type": "gpt_oss",
-  "num_attention_heads": 2,
-  "num_experts_per_tok": 4,
-  "num_hidden_layers": 2,
-  "num_key_value_heads": 1,
-  "num_local_experts": 32,
-  "output_router_logits": false,
-  "pad_token_id": 199999,
-  "quantization_config": {
-    "modules_to_not_convert": null,
-    "quant_method": "mxfp4"
-  },
-  "rms_norm_eps": 1e-05,
-  "rope_scaling": {
-    "beta_fast": 32.0,
-    "beta_slow": 1.0,
-    "factor": 32.0,
-    "original_max_position_embeddings": 4096,
-    "rope_type": "yarn",
-    "truncate": false
-  },
-  "rope_theta": 150000,
-  "router_aux_loss_coef": 0.9,
-  "sliding_window": 128,
-  "swiglu_limit": 7.0,
-  "tie_word_embeddings": true,
-  "transformers_version": "4.56.2",
-  "use_cache": true,
-  "vocab_size": 201088
-}

+version https://git-lfs.github.com/spec/v1
+oid sha256:fdd08f30b14b17a27bb68e5c0fec1291f619140d8724fc14bf7204f105d771d2
+size 1212