Commits · Luigi/VoxSum

fix: reset session data when loading new audio source

0bfe5ff

Luigi commited on Oct 1

feat: allow users to clear speaker names to enable auto-detection

6e6157b

Luigi commited on Oct 1

fix: speaker name propagation in UI after manual edit and auto-detection

89f70aa

Luigi commited on Oct 1

feat: enhance speaker color system to 30 distinct colors

7ac4492

Luigi commited on Oct 1

docs: add comprehensive project summary

e93ffc6

Luigi commited on Oct 1

feat: add custom audio player with visual timeline

2ba9463

Luigi commited on Oct 1

fix: prevent click-to-seek when editing utterance text

4d2f95d

Luigi commited on Oct 1

fix: implement incremental rendering to prevent highlight flicker

f862e7c

Luigi commited on Oct 1

remove unused scripts

cf62bd5

Luigi commited on Oct 1

Consolidate tests under tests/, add LLM default tests with opt-out flag, model selection, README update

913c94a

Luigi commited on Oct 1

add gemma 270m for test use

59519b7

Luigi commited on Sep 27

feat: Improve summarization quality and eliminate repetition

1e9d8a7

Luigi commited on Sep 27

Add cancellation support for long operations

6ca6e79

Luigi commited on Sep 27

Implement multilingual summarization by only touching system prompts

110188a

Luigi commited on Sep 27

remove test scripts

a9ce840

Luigi commited on Sep 27

Fix redundant progress text - simplify UX

ae6b0a5

Luigi commited on Sep 27

Improve UI/UX for slow HuggingFace CPU tier

bcdf9ee

Luigi commited on Sep 27

Complete JavaScript initialization fix

2f0cf23

Luigi commited on Sep 26

Fix marked library loading error

a639fb7

Luigi commited on Sep 26

Add editable speaker names feature

9069ca4

Luigi commited on Sep 26

Add speaker name detection feature

77e98bd

Luigi commited on Sep 26

Fix llama-cpp-python blocking on HF Spaces due to excessive threads

0a56987

Luigi commited on Sep 26

Fix Unicode encoding in export headers for Chinese filenames

81633d5

Luigi commited on Sep 26

Fix filename sanitization for Chinese titles in export

72c7bbc

Luigi commited on Sep 26

Apply zh-cn to zh-tw conversion on generated titles

33359da

Luigi commited on Sep 26

Refine title generation prompt and parameters

ee32282

Luigi commited on Sep 26

Implement document title generation and UI improvements

d8028fb

Luigi commited on Sep 26

Convert summarization.py prompts and comments from French to English

299bf2b

Luigi commited on Sep 26

Clean up legacy Streamlit implementation

228a065

Luigi commited on Sep 26

Add langchain dependency for summarizer module

ca590d4

Luigi commited on Sep 26

gitignore: exclude local development files

17d987e

Luigi commited on Sep 26

cleanup: remove temporary test file

1eed326

Luigi commited on Sep 26

feat: add Markdown rendering support for summaries

b498cb0

Luigi commited on Sep 26

refactor: simplify and robustify summarization module

ae778ed

Luigi commited on Sep 26

fix: handle long transcripts in summarization by chunking

88a0a34

Luigi commited on Sep 26

status in result tab got more informative

030e33b

Luigi commited on Sep 26

update readme

07dbe5e

Luigi commited on Sep 26

add non-tracked

9cd7aca

Luigi commited on Sep 26

update git ignore

da25c8e

Luigi commited on Sep 26

remove models from Dockerfile

c44a627

Luigi commited on Sep 26

remove models/ from .dockerignore

6ee5476

Luigi commited on Sep 26

re-implement in FastAPI

ba4a241

Luigi commited on Sep 26

Fix memory exhaustion on HF Spaces free tier by reducing n_ctx and implementing chunked summarization for long transcripts

09695c9

Luigi commited on Sep 25

remove memory profiler from the proj.

2a2dcfd

Luigi commited on Sep 24

add memory_profiler

e63bfa4

Luigi commited on Sep 24

add profile_diarization.py

cb6b593

Luigi commited on Sep 24

fix: Reduce memory overhead from 2GB→10GB to 2GB→4GB for long audio files

de0b3d5

Luigi commited on Sep 24

feat: CPU-friendly diarization with FAISS

6bf9bbb

Luigi commited on Sep 24

get spk embedding model from hf models

7c59b7a

Luigi commited on Sep 23

bugfix: AttributeError: _ARRAY_API not found et ImportError: numpy.core.multiarray failed to import

41b3df0

Luigi commited on Sep 23

Commit History

fix: reset session data when loading new audio source 0bfe5ff

feat: allow users to clear speaker names to enable auto-detection 6e6157b

fix: speaker name propagation in UI after manual edit and auto-detection 89f70aa

feat: enhance speaker color system to 30 distinct colors 7ac4492

docs: add comprehensive project summary e93ffc6

feat: add custom audio player with visual timeline 2ba9463

fix: prevent click-to-seek when editing utterance text 4d2f95d

fix: implement incremental rendering to prevent highlight flicker f862e7c

remove unused scripts cf62bd5

Consolidate tests under tests/, add LLM default tests with opt-out flag, model selection, README update 913c94a

add gemma 270m for test use 59519b7

feat: Improve summarization quality and eliminate repetition 1e9d8a7

Add cancellation support for long operations 6ca6e79

Implement multilingual summarization by only touching system prompts 110188a

remove test scripts a9ce840

Fix redundant progress text - simplify UX ae6b0a5

Improve UI/UX for slow HuggingFace CPU tier bcdf9ee

Complete JavaScript initialization fix 2f0cf23

Fix marked library loading error a639fb7

Add editable speaker names feature 9069ca4

Add speaker name detection feature 77e98bd

Fix llama-cpp-python blocking on HF Spaces due to excessive threads 0a56987

Fix Unicode encoding in export headers for Chinese filenames 81633d5

Fix filename sanitization for Chinese titles in export 72c7bbc

Apply zh-cn to zh-tw conversion on generated titles 33359da

Refine title generation prompt and parameters ee32282

Implement document title generation and UI improvements d8028fb

Convert summarization.py prompts and comments from French to English 299bf2b

Clean up legacy Streamlit implementation 228a065

Add langchain dependency for summarizer module ca590d4

gitignore: exclude local development files 17d987e

cleanup: remove temporary test file 1eed326

feat: add Markdown rendering support for summaries b498cb0

refactor: simplify and robustify summarization module ae778ed

fix: handle long transcripts in summarization by chunking 88a0a34

status in result tab got more informative 030e33b

update readme 07dbe5e

add non-tracked 9cd7aca

update git ignore da25c8e

remove models from Dockerfile c44a627

remove models/ from .dockerignore 6ee5476

re-implement in FastAPI ba4a241

Fix memory exhaustion on HF Spaces free tier by reducing n_ctx and implementing chunked summarization for long transcripts 09695c9

remove memory profiler from the proj. 2a2dcfd

add memory_profiler e63bfa4

add profile_diarization.py cb6b593

fix: Reduce memory overhead from 2GB→10GB to 2GB→4GB for long audio files de0b3d5

feat: CPU-friendly diarization with FAISS 6bf9bbb

get spk embedding model from hf models 7c59b7a

bugfix: AttributeError: _ARRAY_API not found et ImportError: numpy.core.multiarray failed to import 41b3df0

fix: reset session data when loading new audio source

0bfe5ff

feat: allow users to clear speaker names to enable auto-detection

6e6157b

fix: speaker name propagation in UI after manual edit and auto-detection

89f70aa

feat: enhance speaker color system to 30 distinct colors

7ac4492

docs: add comprehensive project summary

e93ffc6

feat: add custom audio player with visual timeline

2ba9463

fix: prevent click-to-seek when editing utterance text

4d2f95d

fix: implement incremental rendering to prevent highlight flicker

f862e7c

remove unused scripts

cf62bd5

Consolidate tests under tests/, add LLM default tests with opt-out flag, model selection, README update

913c94a

add gemma 270m for test use

59519b7

feat: Improve summarization quality and eliminate repetition

1e9d8a7

Add cancellation support for long operations

6ca6e79

Implement multilingual summarization by only touching system prompts

110188a

remove test scripts

a9ce840

Fix redundant progress text - simplify UX

ae6b0a5

Improve UI/UX for slow HuggingFace CPU tier

bcdf9ee

Complete JavaScript initialization fix

2f0cf23

Fix marked library loading error

a639fb7

Add editable speaker names feature

9069ca4

Add speaker name detection feature

77e98bd

Fix llama-cpp-python blocking on HF Spaces due to excessive threads

0a56987

Fix Unicode encoding in export headers for Chinese filenames

81633d5

Fix filename sanitization for Chinese titles in export

72c7bbc

Apply zh-cn to zh-tw conversion on generated titles

33359da

Refine title generation prompt and parameters

ee32282

Implement document title generation and UI improvements

d8028fb

Convert summarization.py prompts and comments from French to English

299bf2b

Clean up legacy Streamlit implementation

228a065

Add langchain dependency for summarizer module

ca590d4

gitignore: exclude local development files

17d987e

cleanup: remove temporary test file

1eed326

feat: add Markdown rendering support for summaries

b498cb0

refactor: simplify and robustify summarization module

ae778ed

fix: handle long transcripts in summarization by chunking

88a0a34

status in result tab got more informative

030e33b

update readme

07dbe5e

add non-tracked

9cd7aca

update git ignore

da25c8e

remove models from Dockerfile

c44a627

remove models/ from .dockerignore

6ee5476

re-implement in FastAPI

ba4a241

Fix memory exhaustion on HF Spaces free tier by reducing n_ctx and implementing chunked summarization for long transcripts

09695c9

remove memory profiler from the proj.

2a2dcfd

add memory_profiler

e63bfa4

add profile_diarization.py

cb6b593

fix: Reduce memory overhead from 2GB→10GB to 2GB→4GB for long audio files

de0b3d5

feat: CPU-friendly diarization with FAISS

6bf9bbb

get spk embedding model from hf models

7c59b7a

bugfix: AttributeError: _ARRAY_API not found et ImportError: numpy.core.multiarray failed to import

41b3df0