mistralai/Leanstral-1.5-119B-A6B
Full fine-tune vs mistralai/Leanstral-2603. Fine-tune: 100% of the weights changed, across 36 of 36 layers.
Tensors changed: Norms 194/195, Attention 437/636, MLP 1299/2127, other 40/40, Embeddings 1/1, lm_head 1/1
6 derivatives checked
- GZGavinZhao/Leanstral-1.5-119B-A6B-GGUF: LoRA-style
- Abu-Dju/Leanstral-1.5-119B-A6B-GGUF: LoRA-style
- sahilchachra/Leanstral-1.5-119B-A6B-NVFP4: Quantization
- SEBK4C/LEANSTRAL-WHITEDWARF: LoRA-style
- servantofares/Leanstral-1.5-119B-A6B-GGUF: LoRA-style
- mvid/Leanstral-1.5-119B-A6B-MLX-3bit: Quantization
3 other repos hold byte-for-byte identical weight files.