Update README.md
Browse files
README.md
CHANGED
|
@@ -21,6 +21,13 @@ processor = AutoProcessor.from_pretrained(model_id, trust_remote_code=True)
|
|
| 21 |
model = AutoModelForCausalLM.from_pretrained(model_id,trust_remote_code=True)
|
| 22 |
# if have a cuda device
|
| 23 |
model.to('cuda')
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 24 |
```
|
| 25 |
|
| 26 |
# LLM Inference
|
|
@@ -91,6 +98,8 @@ Create a concise caption that accurately describes the main elements in the imag
|
|
| 91 |
"""
|
| 92 |
inputs = processor(text=input_text, images=image, return_tensors="pt")
|
| 93 |
inputs = {key: value.cuda() for key, value in inputs.items()}
|
|
|
|
|
|
|
| 94 |
|
| 95 |
image
|
| 96 |
```
|
|
|
|
| 21 |
model = AutoModelForCausalLM.from_pretrained(model_id,trust_remote_code=True)
|
| 22 |
# if have a cuda device
|
| 23 |
model.to('cuda')
|
| 24 |
+
# else if you have cpu you can use
|
| 25 |
+
model = AutoModelForCausalLM.from_pretrained(
|
| 26 |
+
model_id,
|
| 27 |
+
trust_remote_code=True,
|
| 28 |
+
device_map='cpu' # Assure que le modèle est chargé sur le CPU
|
| 29 |
+
,torch_dtype=torch.bfloat16 # Charger le modèle en demi-précision
|
| 30 |
+
)
|
| 31 |
```
|
| 32 |
|
| 33 |
# LLM Inference
|
|
|
|
| 98 |
"""
|
| 99 |
inputs = processor(text=input_text, images=image, return_tensors="pt")
|
| 100 |
inputs = {key: value.cuda() for key, value in inputs.items()}
|
| 101 |
+
# NB : if you are using bflot16 ==>
|
| 102 |
+
inputs = {key: value.to(dtype=model.dtype) if value.dtype == torch.float32 else value for key, value in inputs.items()}
|
| 103 |
|
| 104 |
image
|
| 105 |
```
|