MCPcopy Create free account

hub / github.com/cwxndl/LLM / functions

Functions215 in github.com/cwxndl/LLM

Method__init__
(self, query, top_k, chunk_size, chunk
chatbot3.0/doc_retrieve.py:26
Method__init__
( self, vocab_size=60930, hidden_size=1600, intermediate_size=14336,
model/moe_config.py:6
Method__init__
( self, vocab_size=60930, hidden_size=1600, intermediate_size=6*1300,
model/model_config.py:6
Method__init__
(self, dim:int, eps:float)
model/modeling_ndl.py:60
Method__init__
(self, dim:int, eps:float)
model/modeling_ndl.py:78
Method__init__
(self, dim, max_position_embeddings: int, base=10000.0, device='cuda',scaling_factor=1.0)
model/modeling_ndl.py:91
Method__init__
(self, config:ndlconfig,layer_idx)
model/modeling_ndl.py:198
Method__init__
(self, config:ndlconfig, layer_idx: int,select_rmsnorm:'NDLRMSNorm')
model/modeling_ndl.py:393
Method__init__
(self, config: ndlconfig)
model/modeling_ndl.py:480
Method__init__
(self, config)
model/modeling_ndl.py:629
Method__init__
(self, dim:int, eps:float)
model/modeling_moe.py:57
Method__init__
(self, dim:int, eps:float)
model/modeling_moe.py:75
Method__init__
(self, dim, max_position_embeddings: int, base=10000.0, device='cuda',scaling_factor=1.0)
model/modeling_moe.py:88
Method__init__
(self, config)
model/modeling_moe.py:184
Method__init__
(self, config)
model/modeling_moe.py:264
Method__init__
(self, config:ndlconfig,layer_idx)
model/modeling_moe.py:338
Method__init__
(self, config:ndlconfig, layer_idx: int,select_rmsnorm:'NDLMOERMSNorm')
model/modeling_moe.py:535
Method__init__
(self, config: ndlconfig)
model/modeling_moe.py:626
Method__init__
(self, config)
model/modeling_moe.py:778
Method__init__
(self, raw_data, tokenizer, max_len)
my_lora/finetune.py:129
Method__len__
(self)
finetune.py:154
Method__len__
(self)
finetune_qwen.py:154
Method__len__
(self)
my_lora/finetune.py:140
Method_get_elements
(self)
chatbot3.0/file_preprocess.py:129
Function_get_unpad_data
(attention_mask)
model/modeling_ndl.py:43
Function_get_unpad_data
(attention_mask)
model/modeling_moe.py:40
Method_init_weights
(self, module)
model/modeling_ndl.py:458
Method_init_weights
(self, module)
model/modeling_moe.py:604
Functionagent
(file, question, top_k, chunk_size, chunk_overlap, llm_api)
chat-bot/app_multi_conversation.py:129
Functionagent
(file, question, top_k, chunk_size, chunk_overlap, llm_api)
chatbot3.0/webui.py:164
Methodbackward
(ctx, grad_output)
model/modeling_moe.py:254
Functionbuild_prompt
(history: list[list[str]])
cli_qwen_demo.py:12
Functionbuild_prompt
(history: list[list[str]])
cli_demo.py:25
Functioncall_llm
(prompt, url)
chat-bot/app_multi_conversation.py:101
Functioncall_llm
(prompt, url)
chatbot3.0/webui.py:136
Functioncircle_print
(total_time: int=60)
cli_qwen_demo.py:20
Functioncircle_print
(total_time: int=60)
cli_demo.py:33
Functionclear_history_
()
chat-bot/app_multi_conversation.py:249
Functionclear_history_
()
chatbot3.0/webui.py:334
Functionclear_iamge_inputs
()
chat-bot/app_multi_conversation.py:254
Functionclear_iamge_inputs
()
chatbot3.0/webui.py:339
Functionclear_inputs
()
chat-bot/app_ndl.py:65
Functionclear_inputs
()
chat-bot/app_multi_conversation.py:242
Functionclear_inputs
()
chat-bot/app_qwen.py:65
Functionclear_inputs
()
chatbot3.0/webui.py:327
Functioncollate_fn
(batch)
ex_pretrain.py:233
Functioncollate_fn
(batch)
finetune.py:80
Functioncollate_fn
(batch)
pretrain.py:232
Functioncollate_fn
(batch)
finetune_qwen.py:81
Functioncollate_fn
(batch)
moe_pretrain.py:150
Functioncollate_fn
(batch)
my_lora/finetune.py:73
Functioncreate_item
(request:Request)
chat-bot/NDLmodel_api.py:18
Functioncreate_item
(request:Request)
chat-bot/qwen2_api.py:20
Functioncreate_item
(request:Request)
chatbot3.0/rerank_api.py:21
Methodforward
(self,x)
model/modeling_ndl.py:73
Methodforward
(self,x)
model/modeling_ndl.py:85
Methodforward
(self ,x, position_ids)
model/modeling_ndl.py:105
Methodforward
(self, x, position_ids)
model/modeling_ndl.py:134
Methodforward
( self, hidden_states: torch.Tensor, #输入attention_block的张量 attention_mask: None, #
model/modeling_ndl.py:233
Methodforward
( self, hidden_states: torch.Tensor, attention_mask: None, position_ids:
model/modeling_ndl.py:321
Methodforward
( self, hidden_states: torch.Tensor, #输入:(batch, seq_len, embed_dim) attention_mask
model/modeling_ndl.py:409
Methodforward
( self, input_ids: torch.LongTensor = None, #input_ids是分词后的token_id,代表了语义信息 attenti
model/modeling_ndl.py:507
Methodforward
( self, input_ids: torch.LongTensor = None, #text2id的token_id序列 attention_mask: Opt
model/modeling_ndl.py:654
Methodforward
(self,x)
model/modeling_moe.py:70
Methodforward
(self,x)
model/modeling_moe.py:82
Methodforward
(self ,x, position_ids)
model/modeling_moe.py:102
Methodforward
(self, x, position_ids)
model/modeling_moe.py:131
Methodforward
(self, hidden_states)
model/modeling_moe.py:201
Methodforward
(ctx, x, loss)
model/modeling_moe.py:246
Methodforward
(self, hidden_states)
model/modeling_moe.py:274
Methodforward
( self, hidden_states: torch.Tensor, #输入attention_block的张量 attention_mask: None, #
model/modeling_moe.py:373
Methodforward
( self, hidden_states: torch.Tensor, attention_mask: None, position_ids:
model/modeling_moe.py:461
Methodforward
( self, hidden_states: torch.Tensor, #输入:(batch, seq_len, embed_dim) attention_mask
model/modeling_moe.py:555
Methodforward
( self, input_ids: torch.LongTensor = None, #input_ids是分词后的token_id,代表了语义信息 attenti
model/modeling_moe.py:656
Methodforward
( self, input_ids: torch.LongTensor = None, #text2id的token_id序列 attention_mask: Opt
model/modeling_moe.py:803
Functiongen_baike
(origin_file)
utils/data_preprocess.py:49
Functiongen_code_github
(input_folder,output_folder)
utils/data_preprocess.py:208
Functiongen_wiki_en
(input_folder, output_folder)
utils/data_preprocess.py:139
Functiongen_wiki_zh
(origin_file, output_file)
utils/data_preprocess.py:168
Methodgenerate
( self, inputs: Optional[torch.Tensor] = None, generation_config: Optional[Generati
model/modeling_moe.py:948
Methodget_decoder
(self)
model/modeling_ndl.py:651
Methodget_decoder
(self)
model/modeling_moe.py:800
Methodget_input_embeddings
(self)
model/modeling_ndl.py:502
Methodget_input_embeddings
(self)
model/modeling_ndl.py:636
Methodget_input_embeddings
(self)
model/modeling_moe.py:651
Methodget_input_embeddings
(self)
model/modeling_moe.py:785
Methodget_output_embeddings
(self)
model/modeling_ndl.py:642
Methodget_output_embeddings
(self)
model/modeling_moe.py:791
Functionmultimode_agent
(file, question)
chat-bot/app_multi_conversation.py:217
Functionmultimode_agent
(file, question)
chatbot3.0/webui.py:302
Methodon_epoch_end
在on_epoch_end时保存一次模型。 TrainingArguments的 save_strategy 中 epoch 和 steps 不兼容。要实现每隔 save_steps 步保存一次检查点,考虑到磁盘空间大小,最多只保存最近3个检查点。
ex_pretrain.py:219
Methodon_epoch_end
在on_epoch_end时保存一次模型。 TrainingArguments的 save_strategy 中 epoch 和 steps 不兼容。要实现每隔 save_steps 步保存一次检查点,考虑到磁盘空间大小,最多只保存最近3个检查点。
finetune.py:66
Methodon_epoch_end
在on_epoch_end时保存一次模型。 TrainingArguments的 save_strategy 中 epoch 和 steps 不兼容。要实现每隔 save_steps 步保存一次检查点,考虑到磁盘空间大小,最多只保存最近3个检查点。
pretrain.py:218
Methodon_epoch_end
在on_epoch_end时保存一次模型。 TrainingArguments的 save_strategy 中 epoch 和 steps 不兼容。要实现每隔 save_steps 步保存一次检查点,考虑到磁盘空间大小,最多只保存最近3个检查点。
finetune_qwen.py:67
Methodon_epoch_end
在on_epoch_end时保存一次模型。 TrainingArguments的 save_strategy 中 epoch 和 steps 不兼容。要实现每隔 save_steps 步保存一次检查点,考虑到磁盘空间大小,最多只保存最近3个检查点。
moe_pretrain.py:136
Methodon_epoch_end
在on_epoch_end时保存一次模型。 TrainingArguments的 save_strategy 中 epoch 和 steps 不兼容。要实现每隔 save_steps 步保存一次检查点,考虑到磁盘空间大小,最多只保存最近3个检查点。
my_lora/finetune.py:59
Methodon_log
在打印 n 次日志后清除cuda缓存,适合低显存设备,能防止OOM
ex_pretrain.py:205
Methodon_log
在打印 n 次日志后清除cuda缓存,适合低显存设备,能防止OOM
finetune.py:52
Methodon_log
在打印 n 次日志后清除cuda缓存,适合低显存设备,能防止OOM
pretrain.py:204
Methodon_log
在打印 n 次日志后清除cuda缓存,适合低显存设备,能防止OOM
finetune_qwen.py:53
← previousnext →101–200 of 215, ranked by callers