city96/llava-llama-3-8b-v1_1-imat-gguf

Same weights, converted (fewer layers) vs xtuner/llava-llama-3-8b-v1_1-transformers. city96/llava-llama-3-8b-v1_1-imat-gguf matches part of xtuner/llava-llama-3-8b-v1_1-transformers (291 of its 686 tensors; the rest of xtuner/llava-llama-3-8b-v1_1-transformers has no counterpart). On the shared part: Same weights stored in another dtype (F16/F32 → F16/F32).

Tensors changed: Embeddings 0/1, Norms 0/65, MLP 0/96, Attention 0/128, lm_head 0/1

On the Hub · Downloads on Model Pulse