Skip to content

use a tinymodel to test generation config which aviod timeout #34482

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

Merged
ArthurZucker merged 2 commits into from
Oct 29, 2024
Merged

use a tinymodel to test generation config which aviod timeout #34482

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
58208b9
use a tinymodel to test generation config which aviod timeout
techkang Oct 29, 2024
3c3ca05
remove tailing whitespace
techkang Oct 29, 2024
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
11 changes: 6 additions & 5 deletions tests/utils/test_modeling_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1544,15 +1544,16 @@ def test_pretrained_low_mem_new_config(self):
self.assertEqual(model.__class__.__name__, model_ref.__class__.__name__)

def test_generation_config_is_loaded_with_model(self):
# Note: `TinyLlama/TinyLlama-1.1B-Chat-v1.0` has a `generation_config.json` containing `max_length: 2048`
# Note: `hf-internal-testing/tiny-random-MistralForCausalLM` has a `generation_config.json`
# containing `bos_token_id: 1`

# 1. Load without further parameters
model = AutoModelForCausalLM.from_pretrained("TinyLlama/TinyLlama-1.1B-Chat-v1.0")
self.assertEqual(model.generation_config.max_length, 2048)
model = AutoModelForCausalLM.from_pretrained(TINY_MISTRAL)
self.assertEqual(model.generation_config.bos_token_id, 1)

# 2. Load with `device_map`
model = AutoModelForCausalLM.from_pretrained("TinyLlama/TinyLlama-1.1B-Chat-v1.0", device_map="auto")
self.assertEqual(model.generation_config.max_length, 2048)
model = AutoModelForCausalLM.from_pretrained(TINY_MISTRAL, device_map="auto")
self.assertEqual(model.generation_config.bos_token_id, 1)

@require_safetensors
def test_safetensors_torch_from_torch(self):
Expand Down