XHToken/Spark-X2.5-1.7B
Full fine-tune vs XHToken/Spark-X2.5-1.7B-Base. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: Embeddings 1/1, Norms 57/57, MLP 84/84, Attention 84/84
18 derivatives checked
- XHToken/Spark-X2.5-1.7B-GGUF: Quantization
- rizzoaiacademy/rizzo-flow: Different architecture
- darioooooo0o/Spark-X2.5-1.7B-Abliterated-GGUF: Quantization
- ngquocvinh/Spark-X2.5-1.7B-GGUF: Quantization
- eaddario/Spark-X2.5-1.7B-GGUF: Same weights, converted
- stornic56/Spark-X2.5-1.7B-GGUF: Same weights, converted
- rizzoaiacademy/rizzo-flow-1.7b: LoRA-style
- darioooooo0o/spark-1.7b-engram: Different architecture
- webbrain-one/webbrain-compass-tiny-xs-v3: LoRA-style
- sizzlebop/Spark-X2.5-1.7B-GGUF: Same weights, converted
- darioooooo0o/Spark-X2.5-1.7B-GGUF: Quantization
- wepiqx/Spark-X2.5-1.7B-ASHQ1-GGUF: Quantization
- SirSahOl/Spark-X2.5-1.7B-chat-mlx-16bit: Same weights, repacked
- OpenIntelligenceNet/Spark-X2.5-1.7B-Uncensored: LoRA-style
- SirSahOl/Spark-X2.5-1.7B-chat-mlx-4bit: Quantization
- ukung/semantic-lite-2: Different architecture
- ZZRI/Feihua-n2-1.7B: LoRA-style
- Lalakai/Hunter-1-1.7B: LoRA-style