Instructions to use peft-internal-testing/tiny-Llama4ForCausalLM with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use peft-internal-testing/tiny-Llama4ForCausalLM with Transformers:

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

pipe = pipeline("text-generation", model="peft-internal-testing/tiny-Llama4ForCausalLM")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("peft-internal-testing/tiny-Llama4ForCausalLM")
model = AutoModelForCausalLM.from_pretrained("peft-internal-testing/tiny-Llama4ForCausalLM", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use peft-internal-testing/tiny-Llama4ForCausalLM with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "peft-internal-testing/tiny-Llama4ForCausalLM"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "peft-internal-testing/tiny-Llama4ForCausalLM",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/peft-internal-testing/tiny-Llama4ForCausalLM

SGLang

How to use peft-internal-testing/tiny-Llama4ForCausalLM 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 "peft-internal-testing/tiny-Llama4ForCausalLM" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "peft-internal-testing/tiny-Llama4ForCausalLM",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "peft-internal-testing/tiny-Llama4ForCausalLM" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "peft-internal-testing/tiny-Llama4ForCausalLM",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use peft-internal-testing/tiny-Llama4ForCausalLM with Docker Model Runner:
```
docker model run hf.co/peft-internal-testing/tiny-Llama4ForCausalLM
```

tiny-Llama4ForCausalLM / config.json

BenjaminB HF Staff

Upload Llama4ForCausalLM

94b6192 verified 8 days ago

Raw History Blame Contribute Delete

1.22 kB

{
  "architectures": [
    "Llama4ForCausalLM"
  ],
  "attention_bias": false,
  "attention_chunk_size": 8192,
  "attention_dropout": 0.0,
  "attn_scale": 0.1,
  "attn_temperature_tuning": true,
  "bos_token_id": 1,
  "dtype": "float32",
  "eos_token_id": 2,
  "floor_scale": 8192,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 8,
  "initializer_range": 0.02,
  "interleave_moe_layer_step": 1,
  "intermediate_size": 32,
  "intermediate_size_mlp": 16384,
  "layer_types": [
    "chunked_attention",
    "chunked_attention"
  ],
  "max_position_embeddings": 131072,
  "model_type": "llama4_text",
  "moe_layers": [
    0,
    1
  ],
  "no_rope_layer_interval": 4,
  "no_rope_layers": [
    1,
    1
  ],
  "num_attention_heads": 4,
  "num_experts_per_tok": 1,
  "num_hidden_layers": 2,
  "num_key_value_heads": 2,
  "num_local_experts": 16,
  "output_router_logits": false,
  "pad_token_id": null,
  "rms_norm_eps": 1e-05,
  "rope_parameters": {
    "rope_theta": 500000,
    "rope_type": "default"
  },
  "router_aux_loss_coef": 0.001,
  "router_jitter_noise": 0.0,
  "tie_word_embeddings": false,
  "transformers_version": "5.21.0.dev0",
  "use_cache": true,
  "use_qk_norm": true,
  "vocab_size": 201135
}

	{
	"architectures": [
	"Llama4ForCausalLM"
	],
	"attention_bias": false,
	"attention_chunk_size": 8192,
	"attention_dropout": 0.0,
	"attn_scale": 0.1,
	"attn_temperature_tuning": true,
	"bos_token_id": 1,
	"dtype": "float32",
	"eos_token_id": 2,
	"floor_scale": 8192,
	"head_dim": 128,
	"hidden_act": "silu",
	"hidden_size": 8,
	"initializer_range": 0.02,
	"interleave_moe_layer_step": 1,
	"intermediate_size": 32,
	"intermediate_size_mlp": 16384,
	"layer_types": [
	"chunked_attention",
	"chunked_attention"
	],
	"max_position_embeddings": 131072,
	"model_type": "llama4_text",
	"moe_layers": [
	0,
	1
	],
	"no_rope_layer_interval": 4,
	"no_rope_layers": [
	1,
	1
	],
	"num_attention_heads": 4,
	"num_experts_per_tok": 1,
	"num_hidden_layers": 2,
	"num_key_value_heads": 2,
	"num_local_experts": 16,
	"output_router_logits": false,
	"pad_token_id": null,
	"rms_norm_eps": 1e-05,
	"rope_parameters": {
	"rope_theta": 500000,
	"rope_type": "default"
	},
	"router_aux_loss_coef": 0.001,
	"router_jitter_noise": 0.0,
	"tie_word_embeddings": false,
	"transformers_version": "5.21.0.dev0",
	"use_cache": true,
	"use_qk_norm": true,
	"vocab_size": 201135
	}