Instructions to use tiny-random/mistral-3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tiny-random/mistral-3 with Transformers:

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

pipe = pipeline("image-text-to-text", model="tiny-random/mistral-3")
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)

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("tiny-random/mistral-3")
model = AutoModelForMultimodalLM.from_pretrained("tiny-random/mistral-3", 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=40)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use tiny-random/mistral-3 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tiny-random/mistral-3"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tiny-random/mistral-3",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/tiny-random/mistral-3

SGLang

How to use tiny-random/mistral-3 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 "tiny-random/mistral-3" \
    --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": "tiny-random/mistral-3",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "tiny-random/mistral-3" \
        --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": "tiny-random/mistral-3",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use tiny-random/mistral-3 with Docker Model Runner:
```
docker model run hf.co/tiny-random/mistral-3
```

mistral-3 / README.md

yujiepan

Upload folder using huggingface_hub

931aa2e verified 8 months ago

preview code

Raw

History Blame Contribute Delete

6.6 kB

	---
	library_name: transformers
	base_model:
	- mistralai/Ministral-3-14B-Reasoning-2512
	---

	This tiny model is intended for debugging. It is randomly initialized using the configuration adapted from [mistralai/Ministral-3-14B-Reasoning-2512](https://huggingface.co/mistralai/Ministral-3-14B-Reasoning-2512).

	### Example usage:

	```python
	from transformers import AutoModel, MistralCommonBackend, Mistral3ForConditionalGeneration
	import torch

	# Load model and tokenizer
	model_id = "tiny-random/mistral-3"
	model = Mistral3ForConditionalGeneration.from_pretrained(
	model_id,
	device_map="cuda",
	torch_dtype="bfloat16",
	trust_remote_code=True,
	)
	tokenizer = MistralCommonBackend.from_pretrained(model_id)
	image_url = "https://static.wikia.nocookie.net/essentialsdocs/images/7/70/Battle.png/revision/latest?cb=20220523172438"
	messages = [
	{
	"role": "user",
	"content": [
	{
	"type": "text",
	"text": "What is this?",
	},
	{"type": "image_url", "image_url": {"url": image_url}},
	],
	},
	]

	tokenized = tokenizer.apply_chat_template(
	messages, return_tensors="pt", return_dict=True)
	tokenized["input_ids"] = tokenized["input_ids"].to(device="cuda")
	tokenized["pixel_values"] = tokenized["pixel_values"].to(
	dtype=torch.bfloat16, device="cuda")
	image_sizes = [tokenized["pixel_values"].shape[-2:]]

	output = model.generate(
	**tokenized.to("cuda"),
	image_sizes=image_sizes,
	max_new_tokens=32,
	)[0]

	decoded_output = tokenizer.decode(output[len(tokenized["input_ids"][0]):])
	print(decoded_output)
	```

	### Codes to create this repo:

	```python
	import json
	from pathlib import Path

	import accelerate
	import torch
	from huggingface_hub import file_exists, hf_hub_download
	from transformers import (
	AutoConfig,
	AutoModelForCausalLM,
	AutoProcessor,
	GenerationConfig,
	set_seed,
	Mistral3ForConditionalGeneration,
	MistralCommonBackend,
	)

	source_model_id = "mistralai/Ministral-3-14B-Reasoning-2512"
	save_folder = "/tmp/tiny-random/mistral-3"

	processor = AutoProcessor.from_pretrained(
	source_model_id, trust_remote_code=True)
	processor.save_pretrained(save_folder)
	processor = MistralCommonBackend.from_pretrained(
	source_model_id, trust_remote_code=True)
	processor.save_pretrained(save_folder)

	with open(hf_hub_download(source_model_id, filename='config.json', repo_type='model'), 'r', encoding='utf-8') as f:
	config_json = json.load(f)
	config_json['text_config'].update({
	"head_dim": 32,
	"hidden_size": 8,
	"intermediate_size": 64,
	"num_attention_heads": 8,
	"num_hidden_layers": 2,
	"num_key_value_heads": 4,
	})
	config_json['vision_config'].update({
	"head_dim": 32,
	"hidden_size": 128,
	"intermediate_size": 128,
	"num_attention_heads": 4,
	"num_hidden_layers": 2,
	})
	with open(f"{save_folder}/config.json", "w", encoding='utf-8') as f:
	json.dump(config_json, f, indent=2)

	config = AutoConfig.from_pretrained(
	save_folder,
	trust_remote_code=True,
	)
	print(config)
	torch.set_default_dtype(torch.bfloat16)
	model = Mistral3ForConditionalGeneration(config)
	torch.set_default_dtype(torch.float32)
	if file_exists(filename="generation_config.json", repo_id=source_model_id, repo_type='model'):
	model.generation_config = GenerationConfig.from_pretrained(
	source_model_id, trust_remote_code=True,
	)
	model.generation_config.do_sample = True
	print(model.generation_config)
	model = model.cpu()
	with torch.no_grad():
	for name, p in sorted(model.named_parameters()):
	torch.nn.init.normal_(p, 0, 0.1)
	print(name, p.shape)
	model.save_pretrained(save_folder)
	print(model)
	```

	### Printing the model:

	```text
	Mistral3ForConditionalGeneration(
	(model): Mistral3Model(
	(vision_tower): PixtralVisionModel(
	(patch_conv): Conv2d(3, 128, kernel_size=(14, 14), stride=(14, 14), bias=False)
	(ln_pre): PixtralRMSNorm((128,), eps=1e-05)
	(transformer): PixtralTransformer(
	(layers): ModuleList(
	(0-1): 2 x PixtralAttentionLayer(
	(attention_norm): PixtralRMSNorm((128,), eps=1e-05)
	(feed_forward): PixtralMLP(
	(gate_proj): Linear(in_features=128, out_features=128, bias=False)
	(up_proj): Linear(in_features=128, out_features=128, bias=False)
	(down_proj): Linear(in_features=128, out_features=128, bias=False)
	(act_fn): SiLUActivation()
	)
	(attention): PixtralAttention(
	(k_proj): Linear(in_features=128, out_features=128, bias=False)
	(v_proj): Linear(in_features=128, out_features=128, bias=False)
	(q_proj): Linear(in_features=128, out_features=128, bias=False)
	(o_proj): Linear(in_features=128, out_features=128, bias=False)
	)
	(ffn_norm): PixtralRMSNorm((128,), eps=1e-05)
	)
	)
	)
	(patch_positional_embedding): PixtralRotaryEmbedding()
	)
	(multi_modal_projector): Mistral3MultiModalProjector(
	(norm): Mistral3RMSNorm((128,), eps=1e-05)
	(patch_merger): Mistral3PatchMerger(
	(merging_layer): Linear(in_features=512, out_features=128, bias=False)
	)
	(linear_1): Linear(in_features=128, out_features=8, bias=False)
	(act): GELUActivation()
	(linear_2): Linear(in_features=8, out_features=8, bias=False)
	)
	(language_model): Ministral3Model(
	(embed_tokens): Embedding(131072, 8, padding_idx=11)
	(layers): ModuleList(
	(0-1): 2 x Ministral3DecoderLayer(
	(self_attn): Ministral3Attention(
	(q_proj): Linear(in_features=8, out_features=256, bias=False)
	(k_proj): Linear(in_features=8, out_features=128, bias=False)
	(v_proj): Linear(in_features=8, out_features=128, bias=False)
	(o_proj): Linear(in_features=256, out_features=8, bias=False)
	)
	(mlp): Ministral3MLP(
	(gate_proj): Linear(in_features=8, out_features=64, bias=False)
	(up_proj): Linear(in_features=8, out_features=64, bias=False)
	(down_proj): Linear(in_features=64, out_features=8, bias=False)
	(act_fn): SiLUActivation()
	)
	(input_layernorm): Ministral3RMSNorm((8,), eps=1e-05)
	(post_attention_layernorm): Ministral3RMSNorm((8,), eps=1e-05)
	)
	)
	(norm): Ministral3RMSNorm((8,), eps=1e-05)
	(rotary_emb): Ministral3RotaryEmbedding()
	)
	)
	(lm_head): Linear(in_features=8, out_features=131072, bias=False)
	)
	```