transformers
106b40c7 - Save dequantized models without the quantizer's weight conversions (#50271)

Commit
18 hours ago
Save dequantized models without the quantizer's weight conversions (#50271) * Save dequantized weights under the model's own names * Drop the quantizer's conversions when a model is dequantized * Apply suggestion from @qgallouedec
Author
Parents
Loading