Modelpublic

simaai/Llama-3.2-3B-Instruct-Autoround-Safetensors

sourceHugging Facellama3.2updated 26d agoView on Hugging Face

0likes176downloads

Model Card

Llama-3.2-3B-Instruct Autoround

Pre-LLiMa Hugging Face checkpoint based on `meta-llama/Llama-3.2-3B-Instruct` at source revision 0cb88a4f764b7a12671c53f0838cd831a0843b95. It remains subject to the base model's license, intended use, and limitations.

Quantization

Component	Method	Weight format	Details
Decoder Linear layers	AutoRound	symmetric INT4, G256	196 targets; `lm_head` excluded
`lm_head`	GPTQ	symmetric INT4, G256	static act-order, block size 128, dampening 0.01
Mixed-precision exceptions	—	source dtype	non-Linear parameters remain at source precision

Calibration used HuggingFaceH4/ultrachat_200k (dataset revision not captured), train_sft[:512], rendered with the source chat template. Token IDs were deterministically concatenated and packed into 512 full 1024-token spans (no shuffle); AutoRound used 200 iterations and batch size 1.

Evaluation

Full wikitext-2-raw-v1 evaluation used the wikitext lm-eval task, no example limit, batch size 1, CUDA, and the same evaluator for source and quantized checkpoints on 2026-07-19.

Checkpoint	Word perplexity	Status
`meta-llama/Llama-3.2-3B-Instruct` source	13.390631	Full run
This UltraChat checkpoint	21.088042	Full run
Absolute degradation	7.697411	Lower is better
Relative degradation	57.484%	`100 * (quantized / source - 1)`

Finite-scale validation and a Transformers chat-generation smoke test passed.

Reproduction

This directory includes the exact quantize.py, recipe.yaml, and versions.txt.

bash

python quantize.py --model-path /path/to/models--meta-llama--Llama-3.2-3B-Instruct --output-dir /path/to/output

Environment

Exact Python, CUDA, Torch, Transformers, llmcompressor, AutoRound, and compressed-tensors versions are recorded in versions.txt.

Deployment

This is the pre-LLiMa quantized Hugging Face artifact. Compile it separately for the target Sima.ai platform and keep compiler output separate.

Limitations

Quantization quality can vary by language, domain, prompt format, context length, and deployment runtime; validate the intended workload independently.