LiquidAI/LFM2.5-Audio-1.5B
Full fine-tune (fewer layers) vs LiquidAI/LFM2-1.2B. LiquidAI/LFM2-1.2B matches part of LiquidAI/LFM2.5-Audio-1.5B (148 of its 931 tensors; the rest of LiquidAI/LFM2.5-Audio-1.5B 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
8 derivatives checked
- LiquidAI/LFM2.5-Audio-1.5B-GGUF: Not comparable yet
- cstr/lfm2-audio-1.5b-GGUF: Not comparable yet
- LiquidAI/LFM2.5-Audio-1.5B-JP: Full fine-tune
- Auguments/LiquidAI-LFM2.5-Audio-1.5B-GGUF-BF16: Not comparable yet
- Paulescu/LFM2.5-Audio-1.5B-OHF-Voice-GGUF: Not comparable yet
- IrieDinamik/LiquidAI-LFM2.5-Audio-1.5B-GGUF: Not comparable yet
- on-device-ml/LFM2.5-Audio-1.5B-OHF-Voice-GGUF: Not comparable yet
- loom-ai-org/lfm2.5-audio-1.5b-tts-loom: Not comparable yet