cleanup: remove mlx-4bit/model.safetensors

cleanup: remove mlx-4bit/config.json

fix: add library_name: mlx for app compatibility

add 8bit: model.safetensors.index.json

add 8bit: config.json

add bf16: model.safetensors.index.json

add bf16: model-00004-of-00004.safetensors

add bf16: model-00003-of-00004.safetensors

add bf16: model-00002-of-00004.safetensors

add bf16: model-00001-of-00004.safetensors

add bf16: config.json

add 5bit: model.safetensors.index.json

add 5bit: model.safetensors

add 5bit: config.json

add 6bit: model.safetensors.index.json

add 6bit: config.json

add 4bit: model.safetensors.index.json

add 4bit: model.safetensors

add 4bit: config.json

add mxfp8: model.safetensors.index.json

add mxfp8: model-00002-of-00002.safetensors

add mxfp8: model-00001-of-00002.safetensors

add mxfp8: config.json

add mxfp4: model.safetensors.index.json

add mxfp4: model.safetensors

add mxfp4: config.json

add nvfp4: model.safetensors.index.json

add nvfp4: model.safetensors

add nvfp4: config.json

add MLX 4bit: tokenizer_config.json

add MLX 4bit: tokenizer.json

add MLX 4bit: model.safetensors.index.json

add MLX 4bit: model.safetensors

add MLX 4bit: generation_config.json

add MLX 4bit: config.json

add MLX 4bit: chat_template.jinja

docs: clean card, cross-link models

docs: remove preliminary benchmarks pending official eval

add processor_config.json from base model

docs: add EUPL-1.2 licence

docs: add benchmarks, clean card, cross-link models

data: Lemma 4bit 5-shot strict pass, correct E4B processor (57.1%)

data: Lemma 4bit 5-shot strict pass, all 14 categories (48.9%)

Fix: add quantized base_model relation for model tree

Add YAML metadata to model card

Add model card

initial commit