SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

[docs] Increase visibility of torch_dtype="auto" (#35067)

Browse Source 
* auto-dtype

* feedback
This commit is contained in:
Steven Liu
2024-12-04 09:18:44 -08:00
committed by GitHub
parent baa3b22137
commit 1ed1de2fec
11 changed files with 49 additions and 35 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
docs/source/en/perf_infer_gpu_one.md
View File

@@ -405,7 +405,7 @@ To load a model in 4-bit for inference, use the `load_in_4bit` parameter. The `d
from transformers import AutoModelForCausalLM
model_name = "bigscience/bloom-2b5"
model_4bit = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto", load_in_4bit=True)
model_4bit = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype="auto", device_map="auto", load_in_4bit=True)
```
To load a model in 4-bit for inference with multiple GPUs, you can control how much GPU RAM you want to allocate to each GPU. For example, to distribute 600MB of memory to the first GPU and 1GB of memory to the second GPU:
@@ -414,7 +414,7 @@ To load a model in 4-bit for inference with multiple GPUs, you can control how m
max_memory_mapping = {0: "600MB", 1: "1GB"}
model_name = "bigscience/bloom-3b"
model_4bit = AutoModelForCausalLM.from_pretrained(
model_name, device_map="auto", load_in_4bit=True, max_memory=max_memory_mapping
model_name, torch_dtype="auto", device_map="auto", load_in_4bit=True, max_memory=max_memory_mapping
)
```
@@ -432,7 +432,7 @@ To load a model in 8-bit for inference, use the `load_in_8bit` parameter. The `d
from transformers import AutoModelForCausalLM, BitsAndBytesConfig
model_name = "bigscience/bloom-2b5"
model_8bit = AutoModelForCausalLM.from_pretrained(model_name, quantization_config=BitsAndBytesConfig(load_in_8bit=True))
model_8bit = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype="auto", quantization_config=BitsAndBytesConfig(load_in_8bit=True))
```
If you're loading a model in 8-bit for text generation, you should use the [`~transformers.GenerationMixin.generate`] method instead of the [`Pipeline`] function which is not optimized for 8-bit models and will be slower. Some sampling strategies, like nucleus sampling, are also not supported by the [`Pipeline`] for 8-bit models. You should also place all inputs on the same device as the model:
@@ -442,7 +442,7 @@ from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
model_name = "bigscience/bloom-2b5"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model_8bit = AutoModelForCausalLM.from_pretrained(model_name, quantization_config=BitsAndBytesConfig(load_in_8bit=True))
model_8bit = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype="auto", quantization_config=BitsAndBytesConfig(load_in_8bit=True))
prompt = "Hello, my llama is cute"
inputs = tokenizer(prompt, return_tensors="pt").to("cuda")
@@ -456,7 +456,7 @@ To load a model in 4-bit for inference with multiple GPUs, you can control how m
max_memory_mapping = {0: "1GB", 1: "2GB"}
model_name = "bigscience/bloom-3b"
model_8bit = AutoModelForCausalLM.from_pretrained(
model_name, device_map="auto", load_in_8bit=True, max_memory=max_memory_mapping
model_name, torch_dtype="auto", device_map="auto", load_in_8bit=True, max_memory=max_memory_mapping
)
```
@@ -515,7 +515,7 @@ quantization_config = BitsAndBytesConfig(
)
tokenizer = AutoTokenizer.from_pretrained("facebook/opt-350m")
model = AutoModelForCausalLM.from_pretrained("facebook/opt-350m", quantization_config=quantization_config)
model = AutoModelForCausalLM.from_pretrained("facebook/opt-350m", torch_dtype="auto", quantization_config=quantization_config)
# enable BetterTransformer
model = model.to_bettertransformer()