Skip to content

Fixed deprecated torch_dtype argument from HuggingFace transformers: replaced with dtype to avoid warning #1152

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
RishabSA wants to merge 1 commit into from
Closed

Fixed deprecated torch_dtype argument from HuggingFace transformers: replaced with dtype to avoid warning #1152

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions demos/LLaMA2_GPU_Quantized.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -406,7 +406,7 @@
},
{
"cell_type": "code",
"execution_count": 7,
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
Expand Down Expand Up @@ -724,7 +724,7 @@
"# inference_dtype = torch.float16\n",
"\n",
"hf_model = AutoModelForCausalLM.from_pretrained(LLAMA_2_7B_CHAT_PATH,\n",
" torch_dtype=inference_dtype,\n",
" dtype=inference_dtype,\n",
" device_map = \"cuda:0\",\n",
" quantization_config=BitsAndBytesConfig(load_in_4bit=True))\n",
"\n",
Expand Down
80 changes: 50 additions & 30 deletions demos/LLaVA.ipynb
View file
Open in desktop

Large diffs are not rendered by default.

2 changes: 1 addition & 1 deletion demos/stable_lm.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -74,7 +74,7 @@
},
{
"cell_type": "code",
"execution_count": 3,
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
Expand Down
2 changes: 1 addition & 1 deletion tests/acceptance/test_hooked_encoder.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -173,7 +173,7 @@ def test_from_pretrained_revision():
@pytest.mark.parametrize("dtype", [torch.bfloat16, torch.float16])
def test_half_precision(dtype):
"""Check the 16 bits loading and inferences."""
model = HookedEncoder.from_pretrained(MODEL_NAME, torch_dtype=dtype)
model = HookedEncoder.from_pretrained(MODEL_NAME, dtype=dtype)
assert model.W_K.dtype == dtype

_ = model(model.tokenizer("Hello, world", return_tensors="pt")["input_ids"])
Expand Down
8 changes: 4 additions & 4 deletions tests/acceptance/test_hooked_transformer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -342,7 +342,7 @@ def benchmark_model_options(

if hf_model is None:
hf_model = AutoModelForCausalLM.from_pretrained(
model_name, torch_dtype=dtype, device_map="auto"
model_name, dtype=dtype, device_map="auto"
)
if tokenizer is None:
tokenizer = AutoTokenizer.from_pretrained(model_name)
Expand Down Expand Up @@ -492,13 +492,13 @@ def check_dtype(dtype, margin, no_processing=False):
for model_path in ["gpt2", "roneneldan/TinyStories-33M", "EleutherAI/pythia-70m"]:
if no_processing:
# For low precision, the processing is not advised.
model = HookedTransformer.from_pretrained_no_processing(model_path, torch_dtype=dtype)
model = HookedTransformer.from_pretrained_no_processing(model_path, dtype=dtype)
else:
model = HookedTransformer.from_pretrained(model_path, torch_dtype=dtype)
model = HookedTransformer.from_pretrained(model_path, dtype=dtype)

hf_model = AutoModelForCausalLM.from_pretrained(
model_path,
torch_dtype=dtype,
dtype=dtype,
).to("cuda" if torch.cuda.is_available() else "cpu")

for layer_name, layer in model.state_dict().items():
Expand Down
10 changes: 5 additions & 5 deletions transformer_lens/loading_from_pretrained.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1903,15 +1903,15 @@ def get_pretrained_state_dict(
hf_model = AutoModelForCausalLM.from_pretrained(
official_model_name,
revision=f"checkpoint-{cfg.checkpoint_value}",
torch_dtype=dtype,
dtype=dtype,
token=huggingface_token if len(huggingface_token) > 0 else None,
**kwargs,
)
elif official_model_name.startswith("EleutherAI/pythia"):
hf_model = AutoModelForCausalLM.from_pretrained(
official_model_name,
revision=f"step{cfg.checkpoint_value}",
torch_dtype=dtype,
dtype=dtype,
token=huggingface_token,
**kwargs,
)
Expand All @@ -1924,21 +1924,21 @@ def get_pretrained_state_dict(
elif "bert" in official_model_name:
hf_model = BertForPreTraining.from_pretrained(
official_model_name,
torch_dtype=dtype,
dtype=dtype,
token=huggingface_token if len(huggingface_token) > 0 else None,
**kwargs,
)
elif "t5" in official_model_name:
hf_model = T5ForConditionalGeneration.from_pretrained(
official_model_name,
torch_dtype=dtype,
dtype=dtype,
token=huggingface_token if len(huggingface_token) > 0 else None,
**kwargs,
)
else:
hf_model = AutoModelForCausalLM.from_pretrained(
official_model_name,
torch_dtype=dtype,
dtype=dtype,
token=huggingface_token if len(huggingface_token) > 0 else None,
**kwargs,
)
Expand Down
Loading