Code
Hub
Workspaces
Following
Trending
Connect
MCP
copy
Create free account
hub
/
github.com/EleutherAI/gpt-neox
/ functions
Functions
652 in github.com/EleutherAI/gpt-neox
⨍
Functions
652
◇
Types & classes
103
Function
test_neoxargs_consume_neox_args
verify megatron args are correctly consumed after sending via deepspeed
tests/neox_args/test_neoxargs_commandline.py:145
Function
test_neoxargs_duplicates
tests that there are no duplicates among parent classes of NeoXArgs
tests/neox_args/test_neoxargs_implementation.py:22
Function
test_neoxargs_fail_instantiate_without_any_params
verify assertion error if required arguments are not provided
tests/neox_args/test_neoxargs_load.py:153
Function
test_neoxargs_fail_instantiate_without_required_params
verify assertion error if required arguments are not provided
tests/neox_args/test_neoxargs_load.py:140
Function
test_neoxargs_load_arguments_13B_local_setup
verify 13B.yml can be loaded without raising validation errors
tests/neox_args/test_neoxargs_load.py:116
Function
test_neoxargs_load_arguments_175B_local_setup
verify 13B.yml can be loaded without raising validation errors
tests/neox_args/test_neoxargs_load.py:132
Function
test_neoxargs_load_arguments_2_7B_local_setup
verify 2-7B.yml can be loaded without raising validation errors
tests/neox_args/test_neoxargs_load.py:100
Function
test_neoxargs_load_arguments_6_7B_local_setup
verify 6-7B.yml can be loaded without raising validation errors
tests/neox_args/test_neoxargs_load.py:108
Function
test_neoxargs_load_arguments_XL_local_setup
verify XL.yml can be loaded without raising validation errors
tests/neox_args/test_neoxargs_load.py:124
Function
test_neoxargs_load_arguments_large_local_setup
verify large.yml can be loaded without raising validation errors
tests/neox_args/test_neoxargs_load.py:92
Function
test_neoxargs_load_arguments_medium_local_setup
verify medium.yml can be loaded without raising validation errors
tests/neox_args/test_neoxargs_load.py:84
Function
test_neoxargs_load_arguments_small_local_setup
verify small.yml can be loaded without raising validation errors
tests/neox_args/test_neoxargs_load.py:66
Function
test_neoxargs_load_arguments_small_local_setup_text_generation
verify small.yml can be loaded together with text generation without raising validation errors
tests/neox_args/test_neoxargs_load.py:74
Function
test_neoxargs_usage
" checks for code pieces of the pattern "args.*" and verifies that such used arg is defined in NeoXArgs
tests/neox_args/test_neoxargs_usage.py:24
Function
test_split_merge
()
tools/merge_mp_partitions.py:137
Function
test_train
(param_dict)
tests/model/test_model_generation.py:69
Function
test_train
(param_dict)
tests/model/test_model_train.py:66
Function
test_train_bf16
(param_dict)
tests/model/test_model_train.py:107
Function
test_train_optimizers
(param_dict)
tests/model/test_model_train.py:132
Method
tok_decode
(self, tokens)
eval_tasks/eval_adapter.py:129
Method
tok_encode
(self, string: str)
eval_tasks/eval_adapter.py:126
Method
tokenize
(self, text)
megatron/tokenizer/tokenizer.py:111
Method
tokenize
(self, text)
megatron/tokenizer/tokenizer.py:174
Method
tokenize
(self, text)
megatron/tokenizer/tokenizer.py:213
Method
tokenize
(self, text: str)
megatron/tokenizer/tokenizer.py:382
Method
tokenize_batch
(self, text_batch: Union[List[str], str])
megatron/tokenizer/tokenizer.py:250
Method
tokenize_batch
(self, text_batch: Union[List[str], str])
megatron/tokenizer/tokenizer.py:295
Method
tokenize_batch
(self, text_batch: Union[List[str], str])
megatron/tokenizer/tokenizer.py:339
Method
tokenize_batch
(self, text_batch: List[str])
megatron/tokenizer/tokenizer.py:385
Method
tokenizer_type
tokenizer type to use when tokenizing data
tools/corpora.py:87
Method
train_mode
Sets up the model for training by turning off k/v caching and setting `parallel output` of the final layer to True, so logits are not
megatron/model/gpt2_model.py:322
Method
urls
URLs from which to download dataset
tools/corpora.py:82
Method
vocab
Dictionary from vocab text token to id token.
megatron/tokenizer/tokenizer.py:100
Method
vocab
(self)
megatron/tokenizer/tokenizer.py:167
Method
vocab
(self)
megatron/tokenizer/tokenizer.py:200
Method
vocab
(self)
megatron/tokenizer/tokenizer.py:240
Method
vocab
(self)
megatron/tokenizer/tokenizer.py:285
Method
vocab
(self)
megatron/tokenizer/tokenizer.py:326
Method
vocab
(self)
megatron/tokenizer/tokenizer.py:374
Method
vocab_file
Vocab file for tokenizer
tools/corpora.py:97
Function
vocab_parallel_cross_entropy
Helper function for the cross entropy.
megatron/mpu/cross_entropy.py:115
Method
vocab_size
(self)
megatron/tokenizer/tokenizer.py:95
Method
vocab_size
(self)
megatron/tokenizer/tokenizer.py:163
Method
vocab_size
(self)
megatron/tokenizer/tokenizer.py:196
Method
vocab_size
(self)
megatron/tokenizer/tokenizer.py:236
Method
vocab_size
(self)
megatron/tokenizer/tokenizer.py:281
Method
vocab_size
(self)
megatron/tokenizer/tokenizer.py:322
Method
vocab_size
(self)
megatron/tokenizer/tokenizer.py:370
Method
vocab_size
(self)
eval_tasks/eval_adapter.py:102
Function
weights_by_num_docs
Builds weights from a multinomial distribution over groups of data according to the number of samples in each group. We sample from a gr
megatron/data/data_utils.py:252
Method
word_embeddings_weight
Easy accessory for the pipeline engine to tie embeddings across stages.
megatron/model/word_embeddings.py:169
Method
write
Write timers to a tensorboard writer
megatron/utils.py:282
← previous
601–652 of 652, ranked by callers