add temperature_length

Author: ngxson

convert_hf_to_gguf.py

@@ -9988,6 +9988,8 @@ def set_vocab(self):
     def set_gguf_parameters(self):
         super().set_gguf_parameters()
         MistralModel.set_mistral_config(self.gguf_writer, self.hparams)
+        yarn_params = self.hparams["yarn"]
+        self.gguf_writer.add_attn_temperature_length(yarn_params["original_max_position_embeddings"])
 
     # TODO @ngxson : this should be in tensor_mapping, but I don't have time for now
     # copied from https://github.com/vllm-project/vllm/blob/main/vllm/model_executor/models/mistral_large_3.py