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

Libraries

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

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

pipe = pipeline("text-generation", model="tiny-random/llama-3")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("tiny-random/llama-3")
model = AutoModelForCausalLM.from_pretrained("tiny-random/llama-3")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.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(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

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

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tiny-random/llama-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/llama-3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

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

SGLang

How to use tiny-random/llama-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/llama-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/llama-3",
		"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 "tiny-random/llama-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/llama-3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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

This tiny model is intended for debugging. It is randomly initialized using the configuration adapted from meta-llama/Llama-3.3-70B-Instruct.

Example usage:

from transformers import pipeline
model_id = "tiny-random/llama-3"
pipe = pipeline(
    "text-generation", model=model_id, device="cuda",
    trust_remote_code=True, max_new_tokens=3,
)
print(pipe("Hello World!"))

Codes to create this repo:

import torch
from transformers import (
    AutoConfig,
    AutoModelForCausalLM,
    AutoTokenizer,
    GenerationConfig,
    pipeline,
    set_seed,
)

source_model_id = "meta-llama/Llama-3.3-70B-Instruct"
save_folder = "/tmp/tiny-random/llama-3"

tokenizer = AutoTokenizer.from_pretrained(
    source_model_id, trust_remote_code=True,
)
tokenizer.save_pretrained(save_folder)

config = AutoConfig.from_pretrained(
    source_model_id, trust_remote_code=True,
)
config.hidden_size = 8
config.intermediate_size = 64
config.num_attention_heads = 16
config.num_key_value_heads = 8
config.head_dim = 32
config.num_hidden_layers = 2

model = AutoModelForCausalLM.from_config(
    config,
    torch_dtype=torch.bfloat16,
    trust_remote_code=True,
)
model.generation_config = GenerationConfig.from_pretrained(
    source_model_id, trust_remote_code=True,
)
set_seed(42)
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)

Printing the model:

LlamaForCausalLM(
  (model): LlamaModel(
    (embed_tokens): Embedding(128256, 8)
    (layers): ModuleList(
      (0-1): 2 x LlamaDecoderLayer(
        (self_attn): LlamaAttention(
          (q_proj): Linear(in_features=8, out_features=512, bias=False)
          (k_proj): Linear(in_features=8, out_features=256, bias=False)
          (v_proj): Linear(in_features=8, out_features=256, bias=False)
          (o_proj): Linear(in_features=512, out_features=8, bias=False)
        )
        (mlp): LlamaMLP(
          (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): LlamaRMSNorm((8,), eps=1e-05)
        (post_attention_layernorm): LlamaRMSNorm((8,), eps=1e-05)
      )
    )
    (norm): LlamaRMSNorm((8,), eps=1e-05)
    (rotary_emb): LlamaRotaryEmbedding()
  )
  (lm_head): Linear(in_features=8, out_features=128256, bias=False)
)