DeepDavid's picture
Upload README.md with huggingface_hub
324be2f verified
|
Raw
History Blame Contribute Delete
1.1 kB
---
license: gemma
base_model: google/gemma-4-e4b-it
tags:
- quantized
- gguf
- candle
- q8_0
---
# gemma-4-e4b-it — Q8_0 weight cache + auxiliary tensors
Quantized deployment artifacts for [google/gemma-4-e4b-it](https://huggingface.co/google/gemma-4-e4b-it).
- `qcache-q8_0.v1.gguf` — GGML **Q8_0** tensors for the model's linear projections,
stored in a GGUF container keyed by the original checkpoint tensor paths
(no metadata KVs). Produced by in-memory quantization of the bf16
checkpoint with a [candle](https://github.com/huggingface/candle)-based
loader.
- `aux-tensors.safetensors` — everything the cache does not carry:
embeddings, norms and other non-projection tensors, in their original
dtypes.
- Config and tokenizer files are copied unmodified from the base
repository.
**Format note**: this is *not* a llama.cpp-compatible GGUF — tensors keep
their original checkpoint names and only linear projections are quantized.
Load it with a runtime that pairs the cache with the auxiliary safetensors.
Weights are redistributed under the same terms as the base model.