Update README.md

README.md

@@ -266,7 +266,7 @@ If your GPU allows, load and run inference in half precision (`torch.float16` or
 
 ```diff
 model = AutoModelForVision2Seq.from_pretrained(
-    "lamm-mit/Cephalo-Idefics-2-vision-8b-beta",
+    "lamm-mit/Cephalo-Idefics-2-vision-10b-alpha",
 +    torch_dtype=torch.float16,    
 ).to(DEVICE)
 ```
@@ -287,7 +287,7 @@ Mke sure to install `flash-attn`. Refer to the [original repository of Flash Att
 
 ```diff
 model = AutoModelForVision2Seq.from_pretrained(
-    "lamm-mit/Cephalo-Idefics-2-vision-8b-beta",
+    "lamm-mit/Cephalo-Idefics-2-vision-10b-alpha",
 +    torch_dtype=torch.bfloat16,    
 +    _attn_implementation="flash_attention_2",
 ).to(DEVICE)
@@ -298,7 +298,7 @@ model = AutoModelForVision2Seq.from_pretrained(
 **4 bit quantization with bitsandbytes**
 
 <details><summary>Click to expand.</summary>
-It is possible to load Idefics2 in 4bits with `bitsandbytes`. Make sure that you have `accelerate` and `bitsandbytes` installed.
+It is possible to load Cephalo-Idefics-2-vision-10b-alpha in 4bits with `bitsandbytes`. Make sure that you have `accelerate` and `bitsandbytes` installed.
 
 ```diff
 + from transformers import BitsAndBytesConfig
@@ -310,7 +310,7 @@ quantization_config = BitsAndBytesConfig(
     bnb_4bit_compute_dtype=torch.bfloat16
 )
 model = AutoModelForVision2Seq.from_pretrained(
-    "lamm-mit/Cephalo-Idefics-2-vision-8b-beta",
+    "lamm-mit/Cephalo-Idefics-2-vision-10b-alpha",
 +    torch_dtype=torch.bfloat16,    
 +    quantization_config=quantization_config,
 ).to(DEVICE)