athirdpath/Cosmos-37b-BASE

Full fine-tune vs athirdpath/LD-Zephyria-37b_HEALED. Fine-tune: 99% of the weights changed, across 95 of 95 layers.

Tensors changed: Embeddings 0/1, Norms 0/191, MLP 285/285, Attention 380/380, lm_head 1/1

2 derivatives checked

• mradermacher/Cosmos-37b-BASE-i1-GGUF: Quantization
• mradermacher/Cosmos-37b-BASE-GGUF: Quantization

On the Hub · Downloads on Model Pulse