Update README.md

README.md

@@ -24,7 +24,7 @@ An open-source LLaMa2 language model of 7b parameters fine-tuned (using as base
 This model is a LLM of 7b parameters based on [NousResearch/Nous-Hermes-llama-2-7b](https://huggingface.co/NousResearch/Nous-Hermes-llama-2-7b), a version of [meta-llama/Llama-2-7b](https://huggingface.co/meta-llama/Llama-2-7b) fine-tuned to follow instructions.
 The model was further fine-tuned in order to follow instructions in italian, using [LoRA](https://arxiv.org/abs/2106.09685) approach and a dataset of 120k random pairs of instruction/answer from [raicrits/Orca_ITA_200k](https://huggingface.co/datasets/raicrits/Orca_ITA_200k).
 
-This repository contains the model weights merged with the LoRA adapters obtained in the fine-tuning procedure (in "float32").
+This repository contains the model weights merged with the LoRA adapters obtained in the fine-tuning procedure.
 
 
 - **Developed by:** Stefano Scotta (stefano.scotta@rai.it)
@@ -77,7 +77,7 @@ model_name = "raicrits/Hermes7b_ITA_v1"
 model = LlamaForCausalLM.from_pretrained(
     model_name,
     device_map="auto",
-#   torch_dtype=torch.bfloat16                #if you want to load quantized model to save GPU memory (it gets only a bit slower)
+    torch_dtype=torch.bfloat16                
 )
 
 tokenizer = AutoTokenizer.from_pretrained("Hermes_ITA_Lora_merged_V2", add_eos_token=False)
@@ -113,6 +113,8 @@ The fine-tuning procedure was done using [LoRA](https://arxiv.org/abs/2106.09685
 
 - learning_rate=2e-4,
 
+- mixed precision training: float16
+
 
  **LoRA configuration:**
 -  r= 8