THUDM/chatglm-6b · Discussions

Update tokenization_chatglm.py

#108 opened 3 months ago by

YiKaiNan

Adding `safetensors` variant of this model

#107 opened 5 months ago by

SFconvertbot

new all pull

#106 opened 5 months ago by

zederer

基于英文语料微调，loss降到多少算是一个比较合理的水平

#105 opened 6 months ago by

TanTaili

ChatGLM3-6B 输出答案包含敏感词解决办法汇总

#104 opened 7 months ago by

731why

ValueError: Tokenizer class ChatGLMTokenizer does not exist or is not currently imported.

#103 opened 7 months ago by

BoHuis

🚩 Report

1

#102 opened 8 months ago by

puffy310

用最新版Transformers报错的解决办法 AttributeError: 'ChatGLMTokenizer' object has no attribute 'sp_tokenizer'. Did you mean: '_tokenize'? return self.sp_tokenizer.num_tokens

3

#101 opened 8 months ago by

PiaoYang

Update tokenization_chatglm.py

2

#100 opened 9 months ago by

puffy310

Problem: query_key_layer_scaling_coeff = float(layer_id + 1)

#99 opened 10 months ago by

Kissacat

accelerate tokenizer

#98 opened 12 months ago by

lugim

chatglm2-6b微调过程中的attention_mask具体是如何实现的？哪一部分是双向，哪一部分是单向？

#97 opened 12 months ago by

someone652314

[AUTOMATED] Model Memory Requirements

#96 opened 12 months ago by

model-sizer-bot

[AUTOMATED] Model Memory Requirements

#95 opened 12 months ago by

model-sizer-bot

Update modeling_chatglm.py

#94 opened about 1 year ago by

Sofianesun

Update tokenization_chatglm.py

#93 opened about 1 year ago by

fengkaige

🚩 Report: Ethical issue(s)

#92 opened about 1 year ago by

syj1314

about self.scaled_masked_softmax in modeling_chatglm.py

#91 opened about 1 year ago by

shl97

🚩 Report : Ethical issue(s)

#89 opened about 1 year ago by

121121wqsas

Request: DOI

#88 opened about 1 year ago by

shili1902

怎么做批量推理或生成？

#86 opened about 1 year ago by

Shaobo1103

Could you support CPU device?

#85 opened about 1 year ago by

changwangss

fix bug in tokenizer `decode` method

#84 opened about 1 year ago by

njunlp

m2pro遇到RuntimeError: Unknown platform: darwin这个是什么问题

#83 opened about 1 year ago by

jindada

Update README.md

#82 opened about 1 year ago by

KaiwenS

多 batch 推理结果不对齐

4

#81 opened about 1 year ago by

shungouxu

local deploy chatglm-6b issues

1

#80 opened about 1 year ago by

guangfan

chatglm-6b在推理的时候可以修改温度吗？

#79 opened about 1 year ago by

Prettyxfan

chatglm-6b如何支持stream流式输出啊

1

#78 opened about 1 year ago by

Zomens

add gradient checkpointing for the final_layernorm module.

#77 opened about 1 year ago by

zhaoqf123

基于chatglm-6b和mnn的离线安卓app

5

#76 opened about 1 year ago by

zhaode

fix: truncate finished output in stream_generate

#74 opened about 1 year ago by

goldenteethCN

KeyError: 'attention_mask'

#73 opened over 1 year ago by

lanshan

v1.1，加载时报错 Unable to load weights from pytorch checkpoint file for '...'. If you tried to load a PyTorch model from a TF 2.0 checkpoint, please set from_tf=True.

3

#72 opened over 1 year ago by

hanswang73

索引超出范围

#71 opened over 1 year ago by

ZXF123

Request: DOI

#70 opened over 1 year ago by

dante314

Error of missing pipeline_tag

1

#69 opened over 1 year ago by

rbai86

求助：跑webDemo抛异常

#68 opened over 1 year ago by

HuangSong

可以用transformers对chatglm-6b做微调吗？

#67 opened over 1 year ago by

lunfan

[BUG/Help] ice_text.model词表长度与config里设置不一致

#65 opened over 1 year ago by

Au3C2

Tokenizer 不支持batch encode

#64 opened over 1 year ago by

next-social

利用chatglm6b进行embedding提取，输入input_ids提取出来的embedding全为0

1

#63 opened over 1 year ago by

XiaoHangjia

怎么用chatglm做embedding

5

#62 opened over 1 year ago by

betterme

chatglm-6B 如何在 Mac M2 上微调

19

#61 opened over 1 year ago by

xiao111

支持batch形式的对话

1

#60 opened over 1 year ago by

curlyfu

输入的最大字符串长度限制

1

#58 opened over 1 year ago by

huvee

🚩 Report

#57 opened over 1 year ago by

YokiLee

Update README.md

#55 opened over 1 year ago by

zyuhang

下面这几个glm微调的方法相同么？

5

#54 opened over 1 year ago by

kewei

When load_in_8bit=True, the chat becomes VERY VERY SLOW and returns nothing

2

#53 opened over 1 year ago by

leoyangsw

Update tokenization_chatglm.py

Adding `safetensors` variant of this model

new all pull

基于英文语料微调，loss降到多少算是一个比较合理的水平

ChatGLM3-6B 输出答案包含 敏感词 解决办法汇总

ValueError: Tokenizer class ChatGLMTokenizer does not exist or is not currently imported.

🚩 Report

用最新版Transformers报错的解决办法 AttributeError: 'ChatGLMTokenizer' object has no attribute 'sp_tokenizer'. Did you mean: '_tokenize'? return self.sp_tokenizer.num_tokens

Update tokenization_chatglm.py

Problem: query_key_layer_scaling_coeff = float(layer_id + 1)

accelerate tokenizer

chatglm2-6b微调过程中的attention_mask具体是如何实现的？哪一部分是双向，哪一部分是单向？

[AUTOMATED] Model Memory Requirements

[AUTOMATED] Model Memory Requirements

Update modeling_chatglm.py

Update tokenization_chatglm.py

🚩 Report: Ethical issue(s)

about self.scaled_masked_softmax in modeling_chatglm.py

🚩 Report : Ethical issue(s)

Request: DOI

怎么做批量推理或生成？

Could you support CPU device?

fix bug in tokenizer `decode` method

m2pro遇到RuntimeError: Unknown platform: darwin这个是什么问题

Update README.md

多 batch 推理结果不对齐

local deploy chatglm-6b issues

chatglm-6b在推理的时候可以修改温度吗？

chatglm-6b如何支持stream流式输出啊

add gradient checkpointing for the final_layernorm module.

基于chatglm-6b和mnn的离线安卓app

fix: truncate finished output in stream_generate

KeyError: 'attention_mask'

v1.1，加载时报错 Unable to load weights from pytorch checkpoint file for '...'. If you tried to load a PyTorch model from a TF 2.0 checkpoint, please set from_tf=True.

索引超出范围

Request: DOI

Error of missing pipeline_tag

求助：跑webDemo抛异常

可以用transformers对chatglm-6b做微调吗？

[BUG/Help] ice_text.model词表长度与config里设置不一致

Tokenizer 不支持batch encode

利用chatglm6b进行embedding提取，输入input_ids提取出来的embedding全为0

怎么用chatglm做embedding

chatglm-6B 如何在 Mac M2 上微调

支持batch形式的对话

输入的最大字符串长度限制

🚩 Report

Update README.md

下面这几个glm微调的方法相同么？

When load_in_8bit=True, the chat becomes VERY VERY SLOW and returns nothing

ChatGLM3-6B 输出答案包含敏感词解决办法汇总