Skip to content

[tests] Adds GPU tests#256

Merged
pacman100 merged 10 commits intohuggingface:mainfrom
younesbelkada:add-gpu-tests
Apr 8, 2023
Merged

[tests] Adds GPU tests#256
pacman100 merged 10 commits intohuggingface:mainfrom
younesbelkada:add-gpu-tests

Conversation

@younesbelkada
Copy link
Copy Markdown
Contributor

@younesbelkada younesbelkada commented Apr 3, 2023

What does this PR do?

Let's continue extensively testing peft !
This PR introduces slow GPU tests that will be run on a custom GCP instance.
For now GPU tests consists of the following

  • basic model loading + inference in 8bit
  • basic model loading + inference on GPU
  • LoRA training in single & multi GPU, using a causalLM & seq2seqLM model, in 8bit

cc @pacman100

@HuggingFaceDocBuilderDev
Copy link
Copy Markdown

HuggingFaceDocBuilderDev commented Apr 3, 2023

The documentation is not available anymore as the PR was closed or merged.

@younesbelkada younesbelkada marked this pull request as ready for review April 7, 2023 10:33
Copy link
Copy Markdown
Contributor

@pacman100 pacman100 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thank you @younesbelkada for adding the much-needed slow tests, essential, LGTM! 🚀🤗

Left a suggestion.

)

model = AutoModelForCausalLM.from_pretrained(self.causal_lm_model_id, device_map="balanced")
tokenizer = AutoTokenizer.from_pretrained(self.seq2seq_model_id)
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
tokenizer = AutoTokenizer.from_pretrained(self.seq2seq_model_id)
tokenizer = AutoTokenizer.from_pretrained(self.causal_lm_model_id)

@pacman100 pacman100 merged commit fdebf8a into huggingface:main Apr 8, 2023
Guy-Bilitski pushed a commit to Guy-Bilitski/peft that referenced this pull request May 13, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants