shisa-ai/shisa-v2.1-qwen3-8b
Full fine-tune vs Qwen/Qwen3-8B. Fine-tune: 100% of the weights changed, across 36 of 36 layers.
Tensors changed: Embeddings 1/1, Norms 133/145, MLP 108/108, Attention 144/144, lm_head 1/1
6 derivatives checked
- mradermacher/shisa-v2.1-qwen3-8b-GGUF: Same weights, converted
- mradermacher/shisa-v2.1-qwen3-8b-i1-GGUF: Quantization
- dahara1/shisa-v2.1-qwen3-8b-UD-japanese-imatrix: Quantization
- XpressAI/shisa-v2.1-qwen3-8b-GGUF: Same weights, converted
- tatsuyaaaaaaa/shisa-v2.1-qwen3-8b-gguf: Same weights, converted
- SiliconJelly/shisa-v2.1-qwen3-8b-UD-japanese-imatrix: Quantization
1 other repos hold byte-for-byte identical weight files.