LiquidAI/d1-omni-600M
Full fine-tune (fewer layers) vs LiquidAI/LFM2.5-Encoder-350M. LiquidAI/LFM2.5-Encoder-350M matches part of LiquidAI/d1-omni-600M (148 of its 1084 tensors; the rest of LiquidAI/d1-omni-600M has no counterpart). On the shared part: Fine-tune: 100% of the weights changed, across 16 of 16 layers.
Tensors changed: Embeddings 2/2, other 30/30, MLP 48/48, Norms 44/44, Attention 24/24
4 derivatives checked
- LiquidAI/d1-omni-600M-GGUF: Not comparable yet
- AtomicChat/d1-omni-600M-GGUF: Not comparable yet
- TechnoBaptist/d1-omni-600M-GGUF: Not comparable yet
- webmp3/Sakura-d1-omni-600M-HighQuality-GGUF: Not comparable yet
1 other repos hold byte-for-byte identical weight files.