Sakalti/SJT-0.5B
Full fine-tune vs Qwen/Qwen2.5-0.5B. 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
9 derivatives checked
- Sakalti/Qwen2.5-1B-Instruct: Same weights, repacked (fewer layers)
- Sakalti/SJT-900M: Partial fine-tune (fewer layers)
- Sakalti/SJTPass-5: Partial fine-tune (fewer layers)
- Sakalti/SJTPass-4: Partial fine-tune (fewer layers)
- mradermacher/SJT-0.5B-GGUF: Same weights, converted
- Sakalti/SakaMoe-3x1.6B-Instruct: MoE built from dense models
- Sakalti/SJTPass-2: Same weights, converted
- Sakalti/SJTpass-1: Same weights, converted (fewer layers)
- Sakalti/SJTpass-3: Partial fine-tune (fewer layers)