Code
Hub
Workspaces
Following
Trending
Connect
MCP
copy
Create free account
hub
/
github.com/JIA-Lab-research/LongLoRA
/ functions
Functions
157 in github.com/JIA-Lab-research/LongLoRA
⨍
Functions
157
◇
Types & classes
28
Method
__init__
(self, name, out_features, drop_path, img_size, pos_type, model_kwargs)
pdf2txt/backbone.py:35
Method
__init__
( self, start_size=4, recent_size=512, k_seq_dim=2, v_seq_dim=2, )
streaming_llm/kv_cache.py:24
Method
__len__
(self)
eval_distributed.py:56
Method
__len__
(self)
supervised-fine-tune-qlora.py:202
Method
__len__
(self)
supervised-fine-tune.py:201
Function
_flash_attn_full
(query, key, value, attention_mask=None, head_mask=None)
gptneox_attn_replace.py:47
Function
_flash_attn_ssa
(query, key, value, attention_mask=None, head_mask=None)
gptneox_attn_replace.py:31
Method
_init_weights
(self, m)
pdf2txt/beit.py:514
Function
_prepare_decoder_attention_mask
( self, attention_mask, input_shape, inputs_embeds, past_key_values_length )
llama_attn_replace.py:344
Function
_prepare_decoder_attention_mask
( self, attention_mask, input_shape, inputs_embeds, past_key_values_length )
llama_attn_replace_sft.py:350
Function
_prepare_decoder_attention_mask_inference
( self, attention_mask, input_shape, inputs_embeds, past_key_values_length )
llama_attn_replace.py:437
Function
_prepare_decoder_attention_mask_inference
( self, attention_mask, input_shape, inputs_embeds, past_key_values_length )
llama_attn_replace_sft.py:443
Method
add
(self, logits: torch.FloatTensor, labels: torch.LongTensor, model_output: object)
eval_distributed.py:134
Function
beit_base_patch16
(pretrained=False, **kwargs)
pdf2txt/beit.py:606
Function
beit_large_patch16
(pretrained=False, **kwargs)
pdf2txt/beit.py:620
Function
build_vit_fpn_backbone
Create a VIT w/ FPN backbone. Args: cfg: a detectron2 CfgNode Returns: backbone (Backbone): backbone module, must be a
pdf2txt/backbone.py:134
Method
compute
(self)
eval_distributed.py:150
Function
dit_base_patch16
(pretrained=False, **kwargs)
pdf2txt/beit.py:634
Function
dit_large_patch16
(pretrained=False, **kwargs)
pdf2txt/beit.py:648
Function
download_url
Downloads the content of an url to a folder. Modified from \ https://github.com/pyg-team/pytorch_geometric/tree/master/torch_geometric A
streaming_llm/utils.py:75
Method
evict_range
(self, past_key_values, start, end)
streaming_llm/kv_cache.py:96
Method
extra_repr
(self)
pdf2txt/beit.py:54
Function
falcon_pos_shift_attention_forward
( self, hidden_states: torch.Tensor, alibi: torch.Tensor, attention_mask: torch.Tensor, la
streaming_llm/pos_shift/modify_falcon.py:19
Method
forward
(self, x)
supervised-fine-tune-qlora.py:330
Method
forward
(self, x)
pdf2txt/beit.py:51
Method
forward
(self, x, rel_pos_bias=None, training_window_size=None)
pdf2txt/beit.py:135
Method
forward
(self, x, rel_pos_bias=None, training_window_size=None)
pdf2txt/beit.py:231
Method
forward
(self, x, position_embedding=None, **kwargs)
pdf2txt/beit.py:262
Method
forward
(self, x)
pdf2txt/beit.py:309
Method
forward
(self, training_window_size)
pdf2txt/beit.py:348
Method
forward
(self, x)
pdf2txt/beit.py:601
Method
forward
Args: x: Tensor of shape (N,C,H,W). H, W must be a multiple of ``self.size_divisibility``. Returns: dict[str
pdf2txt/backbone.py:91
Function
forward_attention
( self, hidden_states: torch.FloatTensor, attention_mask: torch.FloatTensor, p
gptneox_attn_replace.py:85
Function
forward_flashattn
Input shape: Batch x Time x Channel attention_mask: [bsz, q_len]
llama_attn_replace.py:22
Function
forward_flashattn
Input shape: Batch x Time x Channel attention_mask: [bsz, q_len]
llama_attn_replace_sft.py:24
Function
forward_flashattn_full
Input shape: Batch x Time x Channel attention_mask: [bsz, q_len]
llama_attn_replace.py:132
Function
forward_flashattn_full
Input shape: Batch x Time x Channel attention_mask: [bsz, q_len]
llama_attn_replace_sft.py:138
Function
forward_flashattn_inference
( self, hidden_states: torch.Tensor, attention_mask: Optional[torch.Tensor] = None, position_i
llama_attn_replace.py:364
Function
forward_flashattn_inference
( self, hidden_states: torch.Tensor, attention_mask: Optional[torch.Tensor] = None, position_i
llama_attn_replace_sft.py:370
Function
forward_noflashattn
( self, hidden_states: torch.Tensor, attention_mask: Optional[torch.Tensor] = None, position_i
llama_attn_replace.py:225
Function
forward_noflashattn
( self, hidden_states: torch.Tensor, attention_mask: Optional[torch.Tensor] = None, position_i
llama_attn_replace_sft.py:231
Method
get_num_layers
(self)
pdf2txt/beit.py:557
Function
gpt_neox_pos_shift_attention_forward
( self, hidden_states: torch.FloatTensor, attention_mask: torch.FloatTensor, position_ids: tor
streaming_llm/pos_shift/modify_gpt_neox.py:29
Function
llama_pos_shift_attention_forward
( self, hidden_states: torch.Tensor, attention_mask: Optional[torch.Tensor] = None, position_i
streaming_llm/pos_shift/modify_llama.py:36
Function
llama_pos_shift_attention_forward_flashattn
( self, hidden_states: torch.Tensor, attention_mask: Optional[torch.Tensor] = None, position_i
streaming_llm/pos_shift/modify_llama.py:169
Function
load_jsonl
( file_path, )
streaming_llm/utils.py:105
Method
no_weight_decay
(self)
pdf2txt/beit.py:561
Method
output_shape
(self)
pdf2txt/backbone.py:102
Function
parse_args
()
streaming_llm/utils.py:14
Method
rescale
(param, layer_id)
pdf2txt/beit.py:507
Function
response
(prompt)
inference-qlora.py:53
Function
response
(prompt)
inference.py:53
Function
response
(material, question)
demo.py:87
Function
slice1d
(x, start, end)
streaming_llm/kv_cache.py:12
Function
slice2d
(x, start, end)
streaming_llm/kv_cache.py:4
Function
slice3d
(x, start, end)
streaming_llm/kv_cache.py:8
Function
tokenize_fn
(tokenizer, example)
fine-tune.py:93
← previous
101–157 of 157, ranked by callers