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

Remove all traces of low_cpu_mem_usage (#38792)

Browse Source 
* remove it from all py files

* remove it from the doc

* remove it from examples

* style

* remove traces of _fast_init

* Update test_peft_integration.py

* CIs
This commit is contained in:
Cyril Vallez
2025-06-12 16:39:33 +02:00
committed by GitHub
parent 3542e0b844
commit 4b8ec667e9
76 changed files with 100 additions and 598 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
tests/models/helium/test_modeling_helium.py
View File

@@ -87,9 +87,9 @@ class HeliumIntegrationTest(unittest.TestCase):
"Hello, today is a great day to start a new project. I have been working on a new project for a while now and I have"
]
model = AutoModelForCausalLM.from_pretrained(
model_id, low_cpu_mem_usage=True, torch_dtype=torch.bfloat16, revision="refs/pr/1"
).to(torch_device)
model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=torch.bfloat16, revision="refs/pr/1").to(
torch_device
)
tokenizer = AutoTokenizer.from_pretrained(model_id, revision="refs/pr/1")
inputs = tokenizer(self.input_text, return_tensors="pt", padding=True).to(torch_device)