sail/Sailor2-1B
Full fine-tune (fewer layers) vs Qwen/Qwen2.5-0.5B. Qwen/Qwen2.5-0.5B matches part of sail/Sailor2-1B (290 of its 579 tensors; the rest of sail/Sailor2-1B has no counterpart). On the shared part: Fine-tune: 100% of the weights changed, across 24 of 24 layers.
Tensors changed: Embeddings 1/1, Norms 49/49, MLP 72/72, Attention 168/168
13 derivatives checked
- sail/Sailor2-1B-Chat: Full fine-tune
- QuantFactory/Sailor2-1B-Chat-GGUF: Full fine-tune
- mradermacher/Sailor2-1B-GGUF: Same weights, converted
- tensorblock/sail_Sailor2-1B-GGUF: Quantization
- sail/Sailor2-1B-SFT: Full fine-tune
- sail/Sailor2-L-1B: Full fine-tune
- Sakalti/SJT-1B-Alpha: Same weights, converted
- wannaphong/tongyi-model-v1.5-1b-enth: LoRA-style
- wannaphong/tongyi-model-v1.1-1b: LoRA-style
- wannaphong/tongyi-model-v1.3-1b-enth: LoRA-style
- wannaphong/tongyi-model-v1.1-1b-enth: LoRA-style
- wannaphong/tongyi-model-v1.5-1b-th: LoRA-style
- Sakalti/SJT-990M: Same weights, converted