Functions — github.com/AkaliKong/MiniOneRec

CodeHub MCPcopy Create free account

hub / github.com/AkaliKong/MiniOneRec / functions

Functions426 in github.com/AkaliKong/MiniOneRec

⨍ Functions 426◇ Types & classes 56 ↳ Endpoints 24

↓ 1 callersFunctionprocess_dataset_recursive

Process dataset with recursive year reduction like json2csv

data/amazon18_data_process.py:434

↓ 1 callersFunctionresidual_kmeans_constrained

Residual K-means with constrained balanced clustering Args: X: Input data (N, d) K: Number of clusters per level (int or lis

rq/rqkmeans_constrained.py:67

↓ 1 callersFunctionsave_indices_json

(codes, path, use_prefix=True)

rq/rqkmeans_faiss.py:220

↓ 1 callersFunctionset_seed

↓ 1 callersFunctionset_seed

↓ 1 callersFunctionset_seed

ts_rec_sft.py:53

↓ 1 callersFunctionset_seed

↓ 1 callersFunctionset_seed

↓ 1 callersFunctionsetup_seed

↓ 1 callersFunctionsinkhorn_algorithm

(distances, epsilon, sinkhorn_iterations)

rq/models/layers.py:86

↓ 1 callersFunctionsinkhorn_balance_level

(residuals, centroids, capacities=None, *, batch_size=8192, iters=30, tau=None,

rq/rqkmeans_faiss.py:118

↓ 1 callersFunctionsinkhorn_uniform_mapping

(rq, data, codes, *, batch_size=8192, iters=30, tau=None, verbose=True,

rq/rqkmeans_faiss.py:180

↓ 1 callersFunctiontrain_faiss_rq

(data, num_levels=3, codebook_size=256, verbose=True)

rq/rqkmeans_faiss.py:29

↓ 1 callersFunctionunpack_rq_codes

Unpack FAISS's bit-packed codes into integer index arrays Parameters: codes: (N, M_bytes) uint8 array, from rq.compute_codes nbits:

rq/rqkmeans_faiss.py:46

↓ 1 callersFunctionwrite_atomic_files

(args, interaction_list, user2index)

data/amazon23_data_process.py:354

↓ 1 callersMethodz_encode

(self, input_tensor)

(self, input_ids: torch.LongTensor, scores: torch.FloatTensor)

LogitProcessor.py:45

Method__getitem__

ts_rec_data.py:142

Method__getitem__

ts_rec_data.py:276

Method__getitem__

ts_rec_data.py:416

Method__getitem__

ts_rec_data.py:648

Method__getitem__

Method__getitem__

Method__getitem__

rq/datasets.py:33

(self, input_size=64, hidden_size=64, output_size=64, dropout_prob=0.4, device=None)

data_test.py:16

(self, data_path, dataset, index_file=".index.json")

(self, tokenizer)

ts_rec_data.py:15

(self, train_file, tokenizer, max_len=2048, sample=-1, test=False, seed=0, category="", K=4, dedup=False)

ts_rec_data.py:39

Dataset for token2description and description2token tasks. Args: description_file: Path to .index.json file with

ts_rec_data.py:148

Dataset for sid2title and title2sid tasks. Args: item_file: Path to .item.json file with item features

ts_rec_data.py:284

Fusion dataset combining sequence recommendation with item features. Uses semantic IDs for user history, outputs item titles or descr

ts_rec_data.py:424

( self, prefix_allowed_tokens_fn: Callable[[int, torch.Tensor], List[int]], num_beams:

LogitProcessor.py:26

(self, d_in, d_hid, dropout=0.1)

SASRecModules_ori.py:8

(self, data_path, dataset, index_file=".index.json")

ts_rec_sft.py:27

(self, data_path, dataset, index_file=".index.json")

(self, data_source: Sized, repeat_count: int, seed: Optional[int] = None)

minionerec_trainer.py:100

( self, model: Union[str, PreTrainedModel], base_model: str, reward_funcs: Uni

minionerec_trainer.py:211

(self, hidden_size, item_num, state_size, num_filters, filter_sizes, dropout_rate)

(self, hidden_size, item_num, state_size, dropout, device, num_heads=1)

(self, data_df)

(self, tokenizer)

(self, tokenizer=None, max_len=2048, test=False, category="", dedup=False, seed=None)

(self, train_file, sample=-1, seed=0, max_len=2048, category="", dedup=False, tokenizer=None, test=False)

(self, item_file=None, index_file=None, tokenizer=None, max_len=2048, test=False, category="", dedup=False, se

(self, train_file, max_len=2048, sample=-1, seed=0, category="", dedup=False)

(self, train_file, tokenizer, max_len=2048, sample=-1, test = False, seed=0, category="", K=4, dedup=False)

(self, train_file, max_len=2048, sample=-1, seed=0, category="", dedup=False)

(self, train_file, tokenizer, max_len=2048, sample=-1, test=False, seed=0, category="", K=4, dedup=False)

(self, train_file, tokenizer, max_len=2048, sample=-1, test=False, seed=0, category="", K=4, dedup=False)

(self, train_file, tokenizer, max_len=2048, sample=-1, test = False, seed=0, category="", K=4, dedup=False)

Dataset for sid2title and title2sid tasks. Args: item_file: Path to .item.json file with item features

RL-specific dataset for title2sid and description2sid tasks. Returns prompt-completion pairs for RL training. Args:

RL-specific dataset for sequential recommendation using title sequences. Uses user interaction history with item titles to recommend

RL-specific dataset for sid2title tasks. Returns prompt-completion pairs for RL training where input is semantic ID and output is ite

RL-specific dataset for sequential recommendation using semantic IDs in history and outputting item titles. Uses user interaction his

Fusion dataset combining sequence recommendation with item features. Uses semantic IDs for user history, outputs item titles or descr

SFT dataset that uses item titles in user history to predict next item's semantic ID. Args: train_file: Path to

SFT dataset that uses user interaction history and preferences from preference file. Args: user_preference_file:

SFT dataset that uses user interaction history with preferences to predict next item's semantic ID. Uses interaction history from pre

(self, data_path)

rq/datasets.py:8

(self, original_encoder)

rq/generate_indices_plus.py:16

(self, args, model, data_num)

rq/trainer.py:16

(self, original_encoder)

rq/rqkmeans_plus.py:17

(self, n_e, e_dim, beta = 0.25, kmeans_init = False, kmeans_iters = 10, sk_e

rq/models/vq.py:9

(self, n_e_list, e_dim, sk_epsilons, beta = 0.25, kmeans_init = False, kmeans_iters = 100, sk

rq/models/rq.py:13

( self, layers, dropout=0.0, activation="relu", bn=False )

rq/models/layers.py:9

(self, in_dim=768, # num_emb_list=[256,256,256,256], num_em

rq/models/rqvae.py:11

minionerec_trainer.py:109

ts_rec_data.py:52

ts_rec_data.py:207

ts_rec_data.py:349

ts_rec_data.py:532

minionerec_trainer.py:117

rq/datasets.py:38

Function_get_cosine_schedule_with_warmup_lr_lambda

( current_step, *, num_warmup_steps, num_training_steps, num_cycles )

Function_get_cosine_schedule_with_warmup_lr_lambda

( current_step, *, num_warmup_steps, num_training_steps, num_cycles )

ts_rec_sft.py:63

Function_get_cosine_schedule_with_warmup_lr_lambda

( current_step, *, num_warmup_steps, num_training_steps, num_cycles )

Method_get_eval_sampler

(self, eval_dataset)

minionerec_trainer.py:619

Method_get_train_sampler

(self, train_dataset=None)

minionerec_trainer.py:610

Method_set_signature_columns_if_needed

minionerec_trainer.py:594

Function_single_deepseek_request

(model_name, prompt, max_tokens, api_info, base_url, request_id)

rq/text2emb/utils.py:111

Function_single_minimax_request

(model_name, prompt, max_tokens, api_info, base_url, request_id)

rq/text2emb/utils.py:207

Functioncalculate_hit

(sorted_list,topk,true_items,rewards,r_click,total_reward,hit_click,ndcg_click,hit_purchase,ndcg_purchase)

Functioncf_reward

(prompts, completions)

Functioncf_reward

(prompts, completions)

Functioncheck_path

rq/text2emb/utils.py:276

Methodcompute_loss

(self, model, inputs, return_outputs=False, num_items_in_batch=None)

Methodcreate_model_card

Creates a draft of a model card using the information available to the `Trainer`. Args: model_name (`str` or `None`, *op

minionerec_trainer.py:1098

Methoddata_collator

minionerec_trainer.py:343

(self, t: List[int])

ts_rec_data.py:35

(self, t: List[int])

(self, text, bos=False, eos=False)

data_test.py:21

Functionextract_axis_1

(data, indices)

(self, input_tensor)

bias_emb: [batch_size, 1, emb_size]

SASRecModules_ori.py:15

:param queries: A 3d tensor with shape of [N, T_q, C_q] :param keys: A 3d tensor with shape of [N, T_k, C_k] :return

SASRecModules_ori.py:40

← previous next →201–300 of 426, ranked by callers