Sakalti/tara-3.8B

Same weights, converted vs unsloth/Phi-3.5-mini-instruct. Same weights stored in another dtype (BF16 → F16).

Tensors changed: Embeddings 0/1, Norms 0/65, MLP 0/96, Attention 0/128, lm_head 0/1

2 derivatives checked

• mradermacher/tara-3.8B-GGUF: Same weights, converted
• Sakalti/Tara-3.8B-v1.1: Identical copy

2 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse