Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / fkodom/yet-another-retnet issues and pull requests

#27 - An initiallization issue

Issue - State: open - Opened by leor-c 5 months ago - 1 comment

#26 - How do I make a PR?

Issue - State: open - Opened by yunusskeete 6 months ago

#25 - Bug fix: decay mask for bf16, bf32

Pull Request - State: closed - Opened by fkodom 11 months ago

#24 - Some issues regarding _build_decay_mask.

Issue - State: closed - Opened by Doraemonzzz 11 months ago - 3 comments

#23 - How to train with long sequences using chunkwise feature of RetNet?

Issue - State: closed - Opened by calliope-pro 11 months ago - 8 comments

#22 - a more efficient computation of the state in the chunkwise formulation

Pull Request - State: closed - Opened by leor-c 11 months ago

#21 - Slightly more efficient / cleaner implementation of the chunkwise relative pos. enc.

Pull Request - State: closed - Opened by leor-c 11 months ago - 1 comment

#20 - Performance Tuning

Pull Request - State: closed - Opened by fkodom 11 months ago

#19 - CPU benchmark support

Pull Request - State: closed - Opened by fkodom 11 months ago

#18 - Benchmark_inference

Issue - State: closed - Opened by erlebach 11 months ago - 3 comments

#17 - Running benchmark_inference on the CPU

Issue - State: closed - Opened by erlebach 11 months ago - 1 comment

#16 - float32 -> 32-true

Pull Request - State: closed - Opened by fkodom 11 months ago

#15 - Bug Fix: float32 -> 32-true

Pull Request - State: closed - Opened by fkodom 11 months ago

#14 - No [tool.poetry] section in pyproject.toml

Issue - State: closed - Opened by erlebach 11 months ago - 3 comments

#13 - Invalid precision when running train_project_gutenberg

Issue - State: closed - Opened by erlebach 11 months ago - 2 comments

#12 - Change in how input projections are implemented. seem to converge faster

Pull Request - State: open - Opened by draguve about 1 year ago - 1 comment

#11 - Fixed issue where the dimensions of the Group norm seems to be incorrect

Pull Request - State: closed - Opened by draguve about 1 year ago - 2 comments

#9 - Have you ever tried Retnet for vision tasks?

Issue - State: closed - Opened by cnyvfang about 1 year ago - 4 comments

#8 - How's this RetNet useful when throughput is actually lower?

Issue - State: closed - Opened by achen46 about 1 year ago - 2 comments

#7 - bug fix: change F.relu to F.silu

Pull Request - State: closed - Opened by Dongyeongkim about 1 year ago

#6 - About activation function

Issue - State: closed - Opened by Dongyeongkim about 1 year ago - 2 comments

#5 - Example training script

Pull Request - State: closed - Opened by fkodom about 1 year ago

#4 - Chunkwise Formulation

Pull Request - State: closed - Opened by fkodom about 1 year ago

#3 - Changelog of official implementation

Issue - State: closed - Opened by donglixp about 1 year ago - 1 comment

#2 - Update README.md

Pull Request - State: closed - Opened by Amshaker about 1 year ago

#1 - Throughput measurements of parallel and recurrence methods

Issue - State: closed - Opened by Amshaker about 1 year ago - 3 comments