Skip to content
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.

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

StaticLLMPipeline: Support more generation options #1431

Conversation

TolyaTalamanov
Copy link
Collaborator

No description provided.

@github-actions github-actions bot added category: LLM LLM pipeline (stateful, static) category: sampling Sampling / Decoding algorithms labels Dec 24, 2024
@TolyaTalamanov TolyaTalamanov changed the title StaticLLMPipeline: Support multinomial sampling StaticLLMPipeline: Support more generation options Dec 30, 2024
@ilya-lavrenov
Copy link
Contributor

why not to re-use the whole Sampler? Limited to cases when we don't need to fork sequences (like in beam search)

@ilya-lavrenov ilya-lavrenov added this to the 2025.0 milestone Jan 2, 2025
Co-authored-by: Ilya Lavrenov <ilya.lavrenov@intel.com>
@@ -227,7 +227,6 @@ def run_text_generation_genai(input_text, num, model, tokenizer, args, iter_data
gen_config = model.get_generation_config()
gen_config.max_new_tokens = max_gen_tokens
gen_config.num_beams = args["num_beams"]
Copy link
Contributor

@sbalandi sbalandi Jan 2, 2025

Choose a reason for hiding this comment

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

please, remove set_seed(args['seed']) on str 201 and 356 and add gen_config.rng_seed= args["seed"] here and to run_text_generation_genai_with_stream

Copy link
Collaborator Author

Choose a reason for hiding this comment

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

Done, thanks!

Copy link
Contributor

@dmatveev dmatveev left a comment

Choose a reason for hiding this comment

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

I am fine with the change (didn't understand much) but please address @sbalandi 's comment

@TolyaTalamanov TolyaTalamanov added this pull request to the merge queue Jan 4, 2025
Merged via the queue into openvinotoolkit:master with commit 31d632b Jan 4, 2025
59 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
category: continuous batching Continuous batching category: llm_bench Label for tool/llm_bench folder category: LLM LLM pipeline (stateful, static) category: sampling Sampling / Decoding algorithms no-match-files
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants