Ecosyste.ms: Issues

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

GitHub / BlinkDL/ChatRWKV issues and pull requests

#208 - Add more thai tokens to tokenizer

Pull Request - State: open - Opened by wannaphong 3 days ago - 1 comment

#207 - API_DEMO_CHAT.py No module named 'rwkv

Issue - State: open - Opened by zhanghui-china 17 days ago - 1 comment

#206 - applying rwkv6.cu to VisualRWKV-v6

Issue - State: open - Opened by ddddewang0425 2 months ago

#205 - Update model.py

Pull Request - State: open - Opened by AGENDD 3 months ago

#204 - 关于音乐模型的训练方法

Issue - State: open - Opened by momocoQAQ 4 months ago - 1 comment

#203 - B

Issue - State: open - Opened by jewc20009 5 months ago

#202 - 加速模型载入

Issue - State: open - Opened by shouldsee 7 months ago - 2 comments

#200 - 如何选模型基座？

Issue - State: open - Opened by xunyao4dev 11 months ago - 1 comment

#198 - [pip package] Make loading aware that os.environ can change

Issue - State: open - Opened by Maykeye 11 months ago - 2 comments

#197 - add text condition for gen music

Issue - State: open - Opened by jiaqianjing 12 months ago - 1 comment

#196 - model path list

Issue - State: open - Opened by EnricoBeltramo 12 months ago - 1 comment

#195 - mps slower than cpu

Issue - State: open - Opened by fakerybakery 12 months ago - 1 comment

#194 - How to run new v5-Eagle-7B

Issue - State: open - Opened by arun-samespace about 1 year ago - 2 comments

#193 - "cpu fp32i8" strategy not working in RWKV v6 through Python rwkv module

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

#192 - Inference doesn't work on Apple Macbook even when using CPU fp32 as strategy

Issue - State: open - Opened by fredsco about 1 year ago - 1 comment

#191 - eagle-7B

Issue - State: closed - Opened by l0d0v1c about 1 year ago - 1 comment

#189 - 回复总是截断了，如何让回复自然的结束

Issue - State: open - Opened by micronetboy about 1 year ago - 1 comment

#188 - feat: add torch parallel mode for v5.2 model.

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

#187 - 大哥，乱码了

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

#186 - NameError: name 'PIPELINE' is not defined

Issue - State: open - Opened by Siu-Ming over 1 year ago - 1 comment

#185 - Update rwkv5.cu for multiprocessing

Pull Request - State: open - Opened by harrisonvanderbyl over 1 year ago

#184 - fix bf16

Pull Request - State: closed - Opened by daquexian over 1 year ago

#183 - fix multi gpus

Pull Request - State: closed - Opened by daquexian over 1 year ago

#182 - fix jit

Pull Request - State: closed - Opened by daquexian over 1 year ago

#181 - fix dml sampling

Pull Request - State: closed - Opened by daquexian over 1 year ago

#180 - add v5 int8 by unifying float and int8 att

Pull Request - State: closed - Opened by daquexian over 1 year ago

#179 - How to write the RWKV in autogressive style like RNN

Issue - State: open - Opened by HaiFengZeng over 1 year ago - 2 comments

#178 - remove @MyFunction on cuda_att_seq_v5_2

Pull Request - State: closed - Opened by daquexian over 1 year ago

#177 - optimize att_seq_v5_2 by replacing torch.cat with torch.empty

Pull Request - State: closed - Opened by daquexian over 1 year ago

#176 - pure pytorch att_seq_v5_2

Pull Request - State: closed - Opened by daquexian over 1 year ago

#175 - fix kernel compilation on windows

Pull Request - State: closed - Opened by josStorer over 1 year ago

#174 - Support DirectML with minimized code change

Pull Request - State: closed - Opened by pengan1987 over 1 year ago

#173 - RuntimeError: Error building extension 'wkv_cuda_v1'

Issue - State: closed - Opened by scooter99boston over 1 year ago - 2 comments

#172 - [Feature Request] text2music

Issue - State: open - Opened by rayrayraykk over 1 year ago - 2 comments

#171 - Prompt for RAG with RWKV-4-World-7B-v1-20230626-ctx4096

Issue - State: open - Opened by Matthieu-Tinycoaching over 1 year ago - 1 comment

#170 - fix top_p sampling when cumsum exactly equals to top_p

Pull Request - State: closed - Opened by daquexian over 1 year ago

#169 - huggingface无法访问，模型无法下载

Issue - State: closed - Opened by xiaokai-lyk over 1 year ago - 4 comments

#168 - repalce benchmark.py time with tokens/s

Pull Request - State: closed - Opened by BBuf over 1 year ago

#167 - 'No CUDA GPUs are available' in google colab with V100 GPU and high RAM

Issue - State: closed - Opened by uxff over 1 year ago - 2 comments

#167 - 'No CUDA GPUs are available' in google colab with V100 GPU and high RAM

Issue - State: closed - Opened by uxff over 1 year ago - 2 comments

#166 - demo true error ?

Issue - State: open - Opened by malv-c over 1 year ago - 1 comment

#164 - demo ?

Issue - State: closed - Opened by malv-c over 1 year ago - 2 comments

#163 - 开源中文NSFW微调模型

Issue - State: closed - Opened by ZhengJun-AI over 1 year ago

#162 - fix dense_dim error on pytorch 1.13

Pull Request - State: closed - Opened by daquexian over 1 year ago

#161 - add a "stops words" support

Pull Request - State: open - Opened by yynil over 1 year ago

#160 - Add a support to "stop_words" in PIPELINE

Issue - State: open - Opened by yynil over 1 year ago

#158 - support fp16 on v5 models, fix cublas link errors on some platforms

Pull Request - State: closed - Opened by daquexian over 1 year ago

#157 - fast fp16 seq and one mode

Pull Request - State: open - Opened by daquexian over 1 year ago

#156 - fix fp32 run

Pull Request - State: closed - Opened by daquexian over 1 year ago

#155 - fix small model fp16 nan

Pull Request - State: closed - Opened by daquexian over 1 year ago

#154 - [WIP] Add fast cuda kernels for one mode

Pull Request - State: closed - Opened by daquexian over 1 year ago - 1 comment

#152 - feat: prefetching

Pull Request - State: open - Opened by daquexian over 1 year ago

#151 - 很好的模型，但是

Issue - State: open - Opened by dayu1979 over 1 year ago - 4 comments

#150 - Question about the accuracy of smaller checkpoints

Issue - State: open - Opened by avideci over 1 year ago - 2 comments

#149 - cpu运算乱码问题

Issue - State: closed - Opened by lmsdscav over 1 year ago - 1 comment

#148 - "LayerNormKernelImpl" not implemented for 'Half'

Issue - State: closed - Opened by lmsdscav over 1 year ago

#144 - RWKV 模型没有config.josn 文件吗？用text-generation-webui 报错

Issue - State: open - Opened by hopeforus over 1 year ago - 5 comments

#143 - Model produces nonsense output. Problem with tokenization?

Issue - State: closed - Opened by Wehzie over 1 year ago - 1 comment

#142 - Updated utils pipeline

Pull Request - State: closed - Opened by Prunoideae over 1 year ago

#141 - Problem with downloading model weights

Issue - State: closed - Opened by StarDewXXX over 1 year ago - 1 comment

#140 - feat:Implement chat api service

Pull Request - State: open - Opened by asdszg over 1 year ago

#139 - ai会把coc跑团的模组当成游戏的mod而不是coc跑团的模组

Issue - State: closed - Opened by 1luik over 1 year ago - 1 comment

#138 - [Q] End-of-text token disabled

Issue - State: closed - Opened by ellonde over 1 year ago - 3 comments

#137 - Faster tokenizer

Pull Request - State: open - Opened by LoganDark over 1 year ago

#136 - Fix tokenizer under py38.

Pull Request - State: closed - Opened by cryscan over 1 year ago

#135 - 【BUG】about batch generation

Issue - State: open - Opened by irasin over 1 year ago - 1 comment

#134 - How to build mulit-turn training data correctly

Issue - State: closed - Opened by fubincom over 1 year ago - 2 comments

#133 - 斗胆和大佬们讨论一个知识库相关的技术方案是否可行

Issue - State: open - Opened by maxadc over 1 year ago - 2 comments

#132 - pb with tokeniser

Issue - State: closed - Opened by bello7777 over 1 year ago - 1 comment

#131 - 好混乱啊

Issue - State: closed - Opened by thelongestusernameofall over 1 year ago - 2 comments

#130 - foreverDM mods

Pull Request - State: closed - Opened by beingPurple over 1 year ago

#129 - Fit code to the formula notation on paper

Pull Request - State: closed - Opened by kytimmylai over 1 year ago

#128 - 请教关于用custom dataset 微调后的遗忘问题

Issue - State: closed - Opened by valkryhx over 1 year ago - 2 comments

#127 - 输出乱码

Issue - State: closed - Opened by onexuan over 1 year ago - 1 comment

#126 - Raven模型在结合langchain使用时，总是生成对话

Issue - State: closed - Opened by redlion99 over 1 year ago - 2 comments

#125 - foreverDM mods

Pull Request - State: closed - Opened by beingPurple over 1 year ago

#124 - Can TensorRT be used to further accelerate the inference speed of ChatRWKV?

Issue - State: closed - Opened by Pevernow almost 2 years ago - 1 comment

#123 - 中文语料相关

Issue - State: closed - Opened by Joel0o0 almost 2 years ago - 1 comment

#122 - langchian里面 RWKV-v4 的lora怎么增加

Issue - State: closed - Opened by xx-zhang almost 2 years ago - 1 comment

#121 - Update README.md

Pull Request - State: closed - Opened by digger-yu almost 2 years ago

#120 - Not so good...

Issue - State: closed - Opened by grapewheel almost 2 years ago - 1 comment

#119 - 更新一下pip源的rwkv，不是github的最新版本，fp16i8会报错

Issue - State: closed - Opened by StarWorkXc almost 2 years ago - 2 comments

#118 - 干货训练多轮对话系统文档

Issue - State: closed - Opened by gg22mm almost 2 years ago - 8 comments

#117 - BAD!!!

Issue - State: closed - Opened by BeastyZ almost 2 years ago - 1 comment

#116 - readme寫的不清楚

Issue - State: closed - Opened by zzzgit almost 2 years ago - 2 comments

#114 - Inference time for different strategy fp16 and fp16i8

Issue - State: closed - Opened by fubincom almost 2 years ago - 4 comments

#113 - Implement parallel Inference and Generation

Pull Request - State: open - Opened by Nintorac almost 2 years ago

#112 - How to fine tune the pre-trained model on Apple MPS?

Issue - State: open - Opened by elcolie almost 2 years ago - 1 comment

#111 - feat:cpp server

Pull Request - State: open - Opened by ZeldaHuang almost 2 years ago

#110 - Fix tokens for '\n\n' in init prompt.

Pull Request - State: closed - Opened by cryscan almost 2 years ago

#109 - 训练教程

Issue - State: closed - Opened by ZTurboX almost 2 years ago - 1 comment

#108 - auto calc strategy

Pull Request - State: open - Opened by dede874 almost 2 years ago