Ecosyste.ms: Issues

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

GitHub / feifeibear/llmspeculativesampling issues and pull requests

#36 - Deepmind version's probability distribution

Issue - State: open - Opened by YEOMJINSEOP 25 days ago

#33 - Please add a license to this repo

Issue - State: closed - Opened by mbrukman 3 months ago - 1 comment

#32 - another implementation for deepmind version

Pull Request - State: closed - Opened by feifeibear 4 months ago

#32 - another implementation for deepmind version

Pull Request - State: closed - Opened by feifeibear 4 months ago

#31 - Modification about speculative_sampling_v2

Pull Request - State: closed - Opened by sf-Liu 4 months ago - 1 comment

#31 - Modification about speculative_sampling_v2

Pull Request - State: closed - Opened by sf-Liu 4 months ago - 1 comment

#30 - Unable to Reproduce Results from README File

Issue - State: closed - Opened by pppyb 5 months ago - 1 comment

#30 - Unable to Reproduce Results from README File

Issue - State: closed - Opened by pppyb 5 months ago - 1 comment

#29 - Algorithm error about target model sampling in Deepmind's version

Issue - State: closed - Opened by sf-Liu 5 months ago - 10 comments

#28 - 投机采样的速度比target model速度还要慢

Issue - State: closed - Opened by pppyb 5 months ago - 2 comments

#28 - 投机采样的速度比target model速度还要慢

Issue - State: closed - Opened by pppyb 5 months ago - 2 comments

#27 - output logits not match. question about decoding when draft model and target model is the same.

Issue - State: open - Opened by 66RING 8 months ago - 4 comments
Labels: documentation

#26 - KV-cache for Deepmind's speculative sampling

Issue - State: closed - Opened by briancpark 10 months ago - 1 comment

#26 - KV-cache for Deepmind's speculative sampling

Issue - State: closed - Opened by briancpark 10 months ago - 1 comment

#25 - How to use Tensor Core to accelerate Speculative Sampling?

Issue - State: closed - Opened by zhaoyang-star 12 months ago - 1 comment

#25 - How to use Tensor Core to accelerate Speculative Sampling?

Issue - State: closed - Opened by zhaoyang-star 12 months ago - 1 comment

#24 - does it support batchsize > 1

Issue - State: closed - Opened by GuoYi0 about 1 year ago - 2 comments

#23 - 使用微调的llm，耗时增加，无法加速

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

#23 - 使用微调的llm，耗时增加，无法加速

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

#22 - a change in the output of the model

Issue - State: closed - Opened by wenxin-zhu about 1 year ago

#22 - a change in the output of the model

Issue - State: closed - Opened by wenxin-zhu about 1 year ago

#21 - 关于指定device的问题

Issue - State: closed - Opened by pendulum445 about 1 year ago - 3 comments

#21 - 关于指定device的问题

Issue - State: closed - Opened by pendulum445 about 1 year ago - 3 comments

#20 - add share_gpt benchmarking results

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

#20 - add share_gpt benchmarking results

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

#19 - fix random seed setting bugs and support multiple GPUs

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

#19 - fix random seed setting bugs and support multiple GPUs

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

#18 - multi-gpu inference

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

#18 - multi-gpu inference

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

#17 - add profiler

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

#17 - add profiler

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

#16 - llama 1B performance

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

#16 - llama 1B performance

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

#15 - fix global vars bug

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

#15 - fix global vars bug

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

#14 - correct benchmark results display

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

#14 - correct benchmark results display

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

#13 - test llama model

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

#13 - test llama model

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

#12 - test llama model

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

#12 - test llama model

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

#11 - add server

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

#11 - add server

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

#10 - add time benchmarking and organize the directory better

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

#10 - add time benchmarking and organize the directory better

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

#9 - add requirements.txt

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

#9 - add requirements.txt

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

#8 - fix the random seed argument bug

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

#8 - fix the random seed argument bug

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

#7 - Refactor the KV Cache logic

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

#7 - Refactor the KV Cache logic

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

#6 - Add KVCache for the google's version

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

#6 - Add KVCache for the google's version

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

#5 - remove past_key_values usages, because it will lead to wrong answers

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

#5 - remove past_key_values usages, because it will lead to wrong answers

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

#4 - add KVCache optimization for AutoRegressive Sampling

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

#4 - add KVCache optimization for AutoRegressive Sampling

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

#3 - use kv cache for approx model generates

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

#3 - use kv cache for approx model generates

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

#2 - Parallel question

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

#2 - Parallel question

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

#1 - speculative sampling 结果和target model输出结果不一致

Issue - State: closed - Opened by lawo123 over 1 year ago - 5 comments

#1 - speculative sampling 结果和target model输出结果不一致

Issue - State: closed - Opened by lawo123 over 1 year ago - 5 comments