THUDM/chatglm2-6b · Discussions

compatible for transformers >=4.45.0

1

#112 opened 13 days ago by

lqf

Adding `safetensors` variant of this model

#111 opened about 1 month ago by

SFconvertbot

Adding `safetensors` variant of this model

#110 opened 8 months ago by

SFconvertbot

ChatGLM2 failed with bfloat16 using transformers 4.38.2

#109 opened 9 months ago by

jasonjchen

Update tokenization_chatglm.py

#108 opened 10 months ago by

chgk13

Add support for "padding_side" == "right" in ChatGLMTokenizer.

#107 opened 10 months ago by

fengkaige

Fix ignore warning

#105 opened 12 months ago by

callanwu

RuntimeError: nvrtc: error

#104 opened 12 months ago by

yilin510

🚩 Report: Ethical issue(s)

#103 opened 12 months ago by

laozhan

🚩 Report: Ethical issue(s)

1

#101 opened about 1 year ago by

Frank8n

[AUTOMATED] Model Memory Requirements

#100 opened about 1 year ago by

model-sizer-bot

fix tokenizer initialization issue AttributeError: can't set attribute

4

#99 opened about 1 year ago by

katuni4ka

11

#98 opened about 1 year ago by

nicolehuangyx

THUDM/chatglm2-6b accuracy=0 on lambada_openai dataset with lm-eval.

#97 opened about 1 year ago by

changwangss

Failed to load model 'TheBloke • chatglm2 6B q4_0 ggml' Error: llama.cpp: tensor 'tok_embeddings.weight' is missing from model

#96 opened about 1 year ago by

swumagic

model type修改成chatglm2

#94 opened about 1 year ago by

Yuytranvy

Error reported when running code：assert self.padding_side == "left" AssertionError

#91 opened about 1 year ago by

zhangzhang9999

code problem

#90 opened about 1 year ago by

zhangzhang9999

AttributeError: 'BertTokenizerFast' object has no attribute 'build_prompt'

1

#88 opened about 1 year ago by

yunqian2023

AttributeError: 'ChatGLMTokenizer' object has no attribute 'tokenizer'

7

#87 opened about 1 year ago by

mobbsen-aiq

🚩 Report

1

#83 opened about 1 year ago by

boyuaa

RuntimeError: DefaultCPUAllocator: not enough memory: you tried to allocate 224395264 bytes.

#82 opened about 1 year ago by

0xrk

[AUTOMATED] Model Memory Requirements

#81 opened about 1 year ago by

model-sizer-bot

pytorch1.12

1

#80 opened about 1 year ago by

zxyy123

Any support for adding system instructions similar as chatGPT?

#79 opened about 1 year ago by

timpan

apply_query_key_layer_scaling 的用途是什么呢

2

#78 opened about 1 year ago by

yywind

Update tokenizer_config.json

#77 opened over 1 year ago by

LeeSAIF

what is the multi_query_group_num in config.json?

#75 opened over 1 year ago by

lg920810

关于chatglm2-6b进行P-tuning v2出现效果不佳情况的一个记录或讨论

#74 opened over 1 year ago by

LittleGreen

如何在创建模型时就使用int4？

2

#73 opened over 1 year ago by

shamankk

分词器tokenizer错误，无法识别到eos字符

#72 opened over 1 year ago by

zheng-nlper

compatible with DirectML

1

#71 opened over 1 year ago by

davinwang

如何将本地文本文件制作成.bin的模型文件？

1

#70 opened over 1 year ago by

shantone

微调后推理报错 IndexError: piece id is out of range.

4

#69 opened over 1 year ago by

lyy0905

chatgml2_6b

#68 opened over 1 year ago by

dailywsx

fix modeling_chatglm.py error: `-` operator, with a bool tensor is not supported

1

#67 opened over 1 year ago by

shibing624

多卡微调，位置编码显示shape错误，单卡可正常运行

#66 opened over 1 year ago by

smiling-xu

add get_output_embeddings()

#65 opened over 1 year ago by

ranchlai

add get_output_embeddings()

#64 opened over 1 year ago by

ranchlai

prBr

#63 opened over 1 year ago by

JeoHw

🚩 Report

#62 opened over 1 year ago by

fangg2000

Create hh

#60 opened over 1 year ago by

0xMonia

如何拿到输入句子的embedding

2

#59 opened over 1 year ago by

GuiltyInori

尽情->敬请用尽情期待多少感觉有点怪?

#58 opened over 1 year ago by

fj4444

about underscore character

1

#57 opened over 1 year ago by

captainst

You may consider adding `ignore_mismatched_sizes=True` in the model `from_pretrained` method.

#56 opened over 1 year ago by

EthanMiao

模型并行出错并给出修改方案

1

#54 opened over 1 year ago by

yuanzhoulvpi

有无最大生成长度，突然停止生成

3

#53 opened over 1 year ago by

Squidwargg

模型使用ptv2微调后其他能力下降该如何解决

2

#52 opened over 1 year ago by

couldn

🚩 Report

#51 opened over 1 year ago by

leyong

compatible for transformers >=4.45.0

Adding `safetensors` variant of this model

Adding `safetensors` variant of this model

ChatGLM2 failed with bfloat16 using transformers 4.38.2

Update tokenization_chatglm.py

Add support for "padding_side" == "right" in ChatGLMTokenizer.

Fix ignore warning

RuntimeError: nvrtc: error

🚩 Report: Ethical issue(s)

🚩 Report: Ethical issue(s)

[AUTOMATED] Model Memory Requirements

fix tokenizer initialization issue AttributeError: can't set attribute

11

THUDM/chatglm2-6b accuracy=0 on lambada_openai dataset with lm-eval.

Failed to load model 'TheBloke • chatglm2 6B q4_0 ggml' Error: llama.cpp: tensor 'tok_embeddings.weight' is missing from model

model type修改成chatglm2

Error reported when running code：assert self.padding_side == "left" AssertionError

code problem

AttributeError: 'BertTokenizerFast' object has no attribute 'build_prompt'

AttributeError: 'ChatGLMTokenizer' object has no attribute 'tokenizer'

🚩 Report

RuntimeError: DefaultCPUAllocator: not enough memory: you tried to allocate 224395264 bytes.

[AUTOMATED] Model Memory Requirements

pytorch1.12

Any support for adding system instructions similar as chatGPT?

apply_query_key_layer_scaling 的用途是什么呢

Update tokenizer_config.json

what is the multi_query_group_num in config.json?

关于chatglm2-6b进行P-tuning v2出现效果不佳情况的一个记录或讨论

如何在创建模型时就使用int4？

分词器tokenizer错误，无法识别到eos字符

compatible with DirectML

如何将本地文本文件制作成.bin的模型文件？

微调后推理报错 IndexError: piece id is out of range.

chatgml2_6b

fix modeling_chatglm.py error: `-` operator, with a bool tensor is not supported

多卡微调，位置编码显示shape错误，单卡可正常运行

add get_output_embeddings()

add get_output_embeddings()

prBr

🚩 Report

Create hh

如何拿到输入句子的embedding

尽情->敬请 用尽情期待多少感觉有点怪?

about underscore character

You may consider adding `ignore_mismatched_sizes=True` in the model `from_pretrained` method.

模型并行出错并给出修改方案

有无最大生成长度，突然停止生成

模型使用ptv2微调后其他能力下降该如何解决

🚩 Report

尽情->敬请用尽情期待多少感觉有点怪?