Functions — github.com/Tele-AI/Telechat

CodeHub MCPcopy Create free account

hub / github.com/Tele-AI/Telechat / functions

Functions680 in github.com/Tele-AI/Telechat

⨍ Functions 680◇ Types & classes 127↳ Endpoints 2

↓ 1 callersFunctionmark_only_lora_as_trainable

(model: nn.Module, bias: str = 'none')

deepspeed-telechat/utils/module/lora.py:104

↓ 1 callersFunctionmasked_cross_entropy_loss

(logits, labels, loss_mask)

deepspeed-telechat/sft/main.py:210

↓ 1 callersFunctiononly_optimize_lora_parameters

deepspeed-telechat/utils/module/lora.py:168

↓ 1 callersFunctionparse_args

deepspeed-telechat/sft/process_data.py:17

↓ 1 callersFunctionparse_args

deepspeed-telechat/sft/main.py:40

↓ 1 callersFunctionprocess

(id, samples, tokenizer, max_seq_len, num_workers, num_samples, output_path, args)

deepspeed-telechat/utils/data/data_utils.py:108

↓ 1 callersFunctionprocess_concat_data

(text, tokenizer, max_seq_len, args)

deepspeed-telechat/utils/data/data_utils.py:94

↓ 1 callersFunctionrecover_lora

deepspeed-telechat/utils/module/lora.py:149

↓ 1 callersMethodreset_parameters

deepspeed-telechat/utils/module/lora.py:59

↓ 1 callersMethodsplit_tensor_along_last_dim

(self, tensor: torch.Tensor, num_parti

service/vllm_inf/telechat_12B.py:178

↓ 1 callersMethodsplit_tensor_along_last_dim

(self, tensor: torch.Tensor, num_par

models/7B_8bit/modeling_telechat.py:375

↓ 1 callersMethodsplit_tensor_along_last_dim

(self, tensor: torch.Tensor, num_parti

models/12B_4bit/modeling_telechat.py:375

↓ 1 callersMethodsplit_tensor_along_last_dim

(self, tensor: torch.Tensor, num_par

models/7B/modeling_telechat.py:375

↓ 1 callersMethodsplit_tensor_along_last_dim

(self, tensor: torch.Tensor, num_par

models/12B/modeling_telechat.py:375

↓ 1 callersMethodsplit_tensor_along_last_dim

(self, tensor: torch.Tensor, num_par

models/7B_4bit/modeling_telechat.py:375

↓ 1 callersMethodsplit_tensor_along_last_dim

(self, tensor: torch.Tensor, num_parti

models/12B-V2_8bit/modeling_telechat.py:375

↓ 1 callersMethodsplit_tensor_along_last_dim

(self, tensor: torch.Tensor, num_parti

models/12B_8bit/modeling_telechat.py:375

↓ 1 callersMethodsplit_tensor_along_last_dim

(self, tensor: torch.Tensor, num_parti

models/12B-V2/modeling_telechat.py:375

↓ 1 callersFunctionstreamresponse_v2

(tokenizer, query, history, do_sample, max_length, top_k, top_p, temperature, repetition_penalty)

service/telechat_service.py:70

↓ 1 callersFunctiontelechat_gelu_back

gradient of tanh approximation of gelu gradient of actual gelu is: 0.5 * (1. + torch.erf(x * 0.70710678)) + 0.3989423 * x * torch.exp(-0.5

models/7B_8bit/modeling_telechat.py:279

↓ 1 callersFunctiontelechat_gelu_back

gradient of tanh approximation of gelu gradient of actual gelu is: 0.5 * (1. + torch.erf(x * 0.70710678)) + 0.3989423 * x * torch.exp(-0.5 *

models/12B_4bit/modeling_telechat.py:279

↓ 1 callersFunctiontelechat_gelu_back

gradient of tanh approximation of gelu gradient of actual gelu is: 0.5 * (1. + torch.erf(x * 0.70710678)) + 0.3989423 * x * torch.exp(-0.5

models/7B/modeling_telechat.py:279

↓ 1 callersFunctiontelechat_gelu_back

gradient of tanh approximation of gelu gradient of actual gelu is: 0.5 * (1. + torch.erf(x * 0.70710678)) + 0.3989423 * x * torch.exp(-0.5

models/12B/modeling_telechat.py:279

↓ 1 callersFunctiontelechat_gelu_back

gradient of tanh approximation of gelu gradient of actual gelu is: 0.5 * (1. + torch.erf(x * 0.70710678)) + 0.3989423 * x * torch.exp(-0.5

models/7B_4bit/modeling_telechat.py:279

↓ 1 callersFunctiontelechat_gelu_back

gradient of tanh approximation of gelu gradient of actual gelu is: 0.5 * (1. + torch.erf(x * 0.70710678)) + 0.3989423 * x * torch.exp(-0.5 *

models/12B-V2_8bit/modeling_telechat.py:279

↓ 1 callersFunctiontelechat_gelu_back

gradient of tanh approximation of gelu gradient of actual gelu is: 0.5 * (1. + torch.erf(x * 0.70710678)) + 0.3989423 * x * torch.exp(-0.5 *

models/12B_8bit/modeling_telechat.py:279

↓ 1 callersFunctiontelechat_gelu_back

gradient of tanh approximation of gelu gradient of actual gelu is: 0.5 * (1. + torch.erf(x * 0.70710678)) + 0.3989423 * x * torch.exp(-0.5 *

models/12B-V2/modeling_telechat.py:279

↓ 1 callersFunctionto_device

(batch, device)

deepspeed-telechat/utils/utils.py:19

↓ 1 callersMethodtrain

(self, mode=True)

deepspeed-telechat/utils/module/lora.py:56

↓ 1 callersMethodunfuse_lora_weight

deepspeed-telechat/utils/module/lora.py:69

models/7B_8bit/generation_utils.py:53

models/12B_4bit/generation_utils.py:56

models/7B/generation_utils.py:53

models/12B/generation_utils.py:56

models/7B_4bit/generation_utils.py:53

models/12B-V2_8bit/generation_utils.py:56

models/12B_8bit/generation_utils.py:56

models/12B-V2/generation_utils.py:56

Method__deepcopy__

(self, memodict={})

models/7B_8bit/generation_utils.py:58

Method__deepcopy__

(self, memodict={})

models/12B_4bit/generation_utils.py:61

Method__deepcopy__

(self, memodict={})

models/7B/generation_utils.py:58

Method__deepcopy__

(self, memodict={})

models/12B/generation_utils.py:61

Method__deepcopy__

(self, memodict={})

models/7B_4bit/generation_utils.py:58

Method__deepcopy__

(self, memodict={})

models/12B-V2_8bit/generation_utils.py:61

Method__deepcopy__

(self, memodict={})

models/12B_8bit/generation_utils.py:61

Method__deepcopy__

(self, memodict={})

models/12B-V2/generation_utils.py:61

Method__getitem__

deepspeed-telechat/utils/data/data_utils.py:48

Method__getstate__

models/12B_4bit/tokenization_telechat.py:61

Method__getstate__

models/12B/tokenization_telechat.py:61

Method__getstate__

models/12B-V2_8bit/tokenization_telechat.py:61

Method__getstate__

models/12B_8bit/tokenization_telechat.py:61

Method__getstate__

models/12B-V2/tokenization_telechat.py:61

(self, weight, lora_dim=0, lora_scaling=1,

deepspeed-telechat/utils/module/lora.py:15

(self, chosen_dataset)

deepspeed-telechat/utils/data/data_utils.py:40

(self, output_path, seed, dataset_name)

deepspeed-telechat/utils/data/raw_datasets.py:9

(self, output_path, seed, dataset_name)

deepspeed-telechat/utils/data/raw_datasets.py:31

( self, config, hidden_size: int, num_heads: int, num_kv_heads: int,

service/vllm_inf/telechat_12B.py:102

( self, config: PretrainedConfig, cache_config: Optional[CacheConfig] = None,

service/vllm_inf/telechat_12B.py:212

( self, config: PretrainedConfig, cache_config: Optional[CacheConfig] = None,

service/vllm_inf/telechat_12B.py:282

( self, config: PretrainedConfig, cache_config: Optional[CacheConfig] = None,

service/vllm_inf/telechat_12B.py:368

(self, dim, config, base=10000,precision=torch.half)

models/7B_8bit/modeling_telechat.py:84

(self, hidden_size, eps=1e-6)

models/7B_8bit/modeling_telechat.py:143

(self, causal=False, softmax_scale=None, attention_dropout=0.0, device=None, dtype=None)

models/7B_8bit/modeling_telechat.py:168

models/7B_8bit/modeling_telechat.py:319

(self, config: TelechatConfig, layer_idx)

models/7B_8bit/modeling_telechat.py:330

(self, config: TelechatConfig, layer_idx)

models/7B_8bit/modeling_telechat.py:509

(self, *inputs, **kwargs)

models/7B_8bit/modeling_telechat.py:572

(self, config: TelechatConfig)

models/7B_8bit/modeling_telechat.py:597

(self, config: TelechatConfig)

models/7B_8bit/modeling_telechat.py:745

init from a list of dict

models/7B_8bit/generation_utils.py:9

( self, tokenizer, history: History = None, skip_prompt: bool = False, timeout: Optional[float] =

models/7B_8bit/generation_utils.py:69

( self, vocab_size=160256, hidden_size=4096, n_layer=30, n_head=3

models/7B_8bit/configuration_telechat.py:50

( self, vocab_file, unk_token="<unk>", bos_token="<_start>", eos_token

models/12B_4bit/tokenization_telechat.py:25

(self, dim, config, base=10000,precision=torch.half)

models/12B_4bit/modeling_telechat.py:84

(self, hidden_size, eps=1e-6)

models/12B_4bit/modeling_telechat.py:143

(self, causal=False, softmax_scale=None, attention_dropout=0.0, device=None, dtype=None)

models/12B_4bit/modeling_telechat.py:168

models/12B_4bit/modeling_telechat.py:319

(self, config: TelechatConfig, layer_idx)

models/12B_4bit/modeling_telechat.py:330

(self, config: TelechatConfig, layer_idx)

models/12B_4bit/modeling_telechat.py:509

(self, *inputs, **kwargs)

models/12B_4bit/modeling_telechat.py:572

(self, config: TelechatConfig)

models/12B_4bit/modeling_telechat.py:597

(self, config: TelechatConfig)

models/12B_4bit/modeling_telechat.py:745

init from a list of dict

models/12B_4bit/generation_utils.py:9

( self, tokenizer, history: History = None, skip_prompt: bool = False, timeout: Optional[float] =

models/12B_4bit/generation_utils.py:72

( self, vocab_size=160256, hidden_size=4096, n_layer=30, n_head=32,

models/12B_4bit/configuration_telechat.py:53

(self, dim, config, base=10000,precision=torch.half)

models/7B/modeling_telechat.py:84

(self, hidden_size, eps=1e-6)

models/7B/modeling_telechat.py:143

(self, causal=False, softmax_scale=None, attention_dropout=0.0, device=None, dtype=None)

models/7B/modeling_telechat.py:168

models/7B/modeling_telechat.py:319

(self, config: TelechatConfig, layer_idx)

models/7B/modeling_telechat.py:330

(self, config: TelechatConfig, layer_idx)

models/7B/modeling_telechat.py:509

(self, *inputs, **kwargs)

models/7B/modeling_telechat.py:572

(self, config: TelechatConfig)

models/7B/modeling_telechat.py:597

(self, config: TelechatConfig)

models/7B/modeling_telechat.py:745

init from a list of dict

models/7B/generation_utils.py:9

( self, tokenizer, history: History = None, skip_prompt: bool = False, timeout: Optional[float] =

models/7B/generation_utils.py:69

( self, vocab_size=160256, hidden_size=4096, n_layer=30, n_head=3

models/7B/configuration_telechat.py:50

( self, vocab_file, unk_token="<unk>", bos_token="<_start>", eos_token

models/12B/tokenization_telechat.py:25

(self, dim, config, base=10000,precision=torch.half)

models/12B/modeling_telechat.py:84

(self, hidden_size, eps=1e-6)

models/12B/modeling_telechat.py:143

← previous next →201–300 of 680, ranked by callers