LLM Speech Summarization#418

Open

eric-mccann-pro wants to merge 178 commits intodevelopfrom

feat/qwen-speech-summarization

Commits on Dec 15, 2025

Mock an LLM generator's events stream. Run pytest if RUN_TESTS is true
eric-mccann-pro
committed

Commits on Dec 16, 2025

Commits on Dec 17, 2025

Commits on Dec 24, 2025

Fix path to vllm-entrypoint.sh.
jrobble
committed

Commits on Jan 2, 2026

Commits on Jan 15, 2026

Commits on Jan 24, 2026

Use HF_HUB_OFFLINE before import.
jrobble
committed

Commits on Jan 30, 2026

Fix bug reading classifiers.json from $MPF_HOME.
jrobble
committed
Add a todo about manually aglommerating lists instead of making the model do it with summary_of_summaries
eric-mccann-pro
committed
Force minimum_chunk_size=2 by default to force json array chunking to either make progress or error RATHER THAN possibly iterating forever in summarize_summaries if only a single summary can be sum…
eric-mccann-pro
committed
descriptor: prevent segmentation to avoid redundant summarization between workers
eric-mccann-pro
committed
Remove vestigial functions from non-component qwen summarization implementation
eric-mccann-pro
committed
Replace deprecated pkg_resources
eric-mccann-pro
committed
Merge remote-tracking branch 'origin/feat/qwen-speech-summarization' into feat/qwen-speech-summarization
eric-mccann-pro
committed
Use PROMPT_TEMPLATE property default to handle both the overriden and default cases of jinja template loading
eric-mccann-pro
committed
_underscore_private_class_methods
eric-mccann-pro
committed
Fix test_slapchop test expected to match minimum 2-object chunk size for chunking json objects
eric-mccann-pro
committed
Assume HF_HUB_OFFLINE shouldn't be one for pytest, allowing downloading the tokenizer in a development context
eric-mccann-pro
committed
Add a shortcut for the tail of summarize_summaries grouping. Don't summarize unary chunks as they cannot be summarized further
eric-mccann-pro
committed
Revert 39810e4 and 5683ff9: min_grouping is selected in summarize_summaries. We don't need to change the default for min_grouping on split_array_into_chunks
eric-mccann-pro
committed
If chunk isn't a string, don't loads it
eric-mccann-pro
committed
pydantic: fix deprecated call to .json()
eric-mccann-pro
committed
Fixup path handling for prompt_template AND its default value
eric-mccann-pro
committed
CORRECT the extension on the prompt template
eric-mccann-pro
committed
If you're a float, be a float
eric-mccann-pro
committed
Correct property name
eric-mccann-pro
committed

Commits on Feb 2, 2026

Copy topics+entities to main detection more defensively
eric-mccann-pro
committed

Commits on Feb 5, 2026

Update python/QwenSpeechSummarization/qwen_speech_summarization_component/qwen_speech_summarization_component.py

eric-mccann-pro
and
tstrass
authored

Commits on Feb 11, 2026

Commits on Feb 16, 2026

Commits on Feb 24, 2026

Commits on Feb 25, 2026

Commits on Feb 27, 2026

Commits on Mar 4, 2026

Commits on Mar 6, 2026

Commits on Mar 9, 2026

Use TEXT if no TRANSCRIPT and no TRANSLATION
eric-mccann-pro
committed

Commits on Mar 12, 2026

llmspeechsummarization/Dockerfile: download both qwen and gpt-oss tokenizers during docker build
eric-mccann-pro
committed

Commits on Mar 20, 2026

Merge branch 'develop' into feat/qwen-speech-summarization
jrobble
committed

Commits on Mar 24, 2026

Merge remote-tracking branch 'origin/develop' into feat/qwen-speech-summarization
tstrass
committed