Instructions to use open-machine/Llama-3.2-1B-FlashNorm with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use open-machine/Llama-3.2-1B-FlashNorm with Transformers:

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

pipe = pipeline("text-generation", model="open-machine/Llama-3.2-1B-FlashNorm")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("open-machine/Llama-3.2-1B-FlashNorm")
model = AutoModelForCausalLM.from_pretrained("open-machine/Llama-3.2-1B-FlashNorm")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use open-machine/Llama-3.2-1B-FlashNorm with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "open-machine/Llama-3.2-1B-FlashNorm"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "open-machine/Llama-3.2-1B-FlashNorm",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/open-machine/Llama-3.2-1B-FlashNorm

SGLang

How to use open-machine/Llama-3.2-1B-FlashNorm 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 "open-machine/Llama-3.2-1B-FlashNorm" \
    --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": "open-machine/Llama-3.2-1B-FlashNorm",
		"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 "open-machine/Llama-3.2-1B-FlashNorm" \
        --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": "open-machine/Llama-3.2-1B-FlashNorm",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use open-machine/Llama-3.2-1B-FlashNorm with Docker Model Runner:
```
docker model run hf.co/open-machine/Llama-3.2-1B-FlashNorm
```

nielsr HF Staff commited on Apr 23

Commit

ef5a0e4

verified ·

1 Parent(s): c7abffa

Add library_name and update model card

Browse files

This PR adds `library_name: transformers` to the YAML metadata to enable the library's automated code snippets on the hub. It also includes a link to the paper [FlashNorm: Fast Normalization for Transformers](https://huggingface.co/papers/2407.09577) in the model description.

Files changed (1) hide show

README.md +11 -8

README.md CHANGED Viewed

@@ -1,17 +1,20 @@
 ---
-license: llama3.2
 base_model: meta-llama/Llama-3.2-1B
-tags:
-  - flashnorm
-  - transformer-tricks
-  - efficient-inference
-  - weightless-rmsnorm
 pipeline_tag: text-generation
 ---
 # Llama-3.2-1B-FlashNorm
-FlashNorm-prepared checkpoint of [meta-llama/Llama-3.2-1B](https://huggingface.co/meta-llama/Llama-3.2-1B). Mathematically equivalent to the source model. The per-channel RMSNorm weight tensors (`input_layernorm.weight`, `post_attention_layernorm.weight`, `model.norm.weight`) are folded into the following linear layers and then removed from the state dict entirely.
 > **Framework support note.** Stock vLLM currently does not load this checkpoint because the norm weight tensors are absent. The upstream patch to accept missing tensors is tracked at: **TBD (vLLM issue link)**. Until the patch lands, use HuggingFace Transformers; it loads this with a warning that norm weights were not initialized and defaults them to ones, which is the correct behavior for FlashNorm.
@@ -55,4 +58,4 @@ Not yet supported. See the tracking issue linked above.
 ## License
-Inherited from the source model.

 ---
 base_model: meta-llama/Llama-3.2-1B
+library_name: transformers
+license: llama3.2
 pipeline_tag: text-generation
+tags:
+- flashnorm
+- transformer-tricks
+- efficient-inference
+- weightless-rmsnorm
 ---
 # Llama-3.2-1B-FlashNorm
+FlashNorm-prepared checkpoint of [meta-llama/Llama-3.2-1B](https://huggingface.co/meta-llama/Llama-3.2-1B). This model was presented in the paper [FlashNorm: Fast Normalization for Transformers](https://huggingface.co/papers/2407.09577).
+Mathematically equivalent to the source model. The per-channel RMSNorm weight tensors (`input_layernorm.weight`, `post_attention_layernorm.weight`, `model.norm.weight`) are folded into the following linear layers and then removed from the state dict entirely.
 > **Framework support note.** Stock vLLM currently does not load this checkpoint because the norm weight tensors are absent. The upstream patch to accept missing tensors is tracked at: **TBD (vLLM issue link)**. Until the patch lands, use HuggingFace Transformers; it loads this with a warning that norm weights were not initialized and defaults them to ones, which is the correct behavior for FlashNorm.
 ## License
+Inherited from the source model.