feat(data): walk folder to get dataset_type_ids_map (#477)

* walk folder to get dataset_type_ids_map

* fix a bug

internlm/data/packed_dataset.py

@@ -15,7 +15,7 @@ from tqdm import tqdm
 
 from internlm.core.context import global_context as gpc
 from internlm.data.single_dataset import JsonlDataset
-from internlm.data.utils import get_dataset_type_id
+from internlm.data.utils import get_dataset_type_id, get_dataset_type_ids_map
 from internlm.utils.logger import get_logger
 
 DEFAULT_SEED = 1024
@@ -373,6 +373,8 @@ def get_packed_dataset_without_short_length(
     datasets = []
     delete_samples = 0
 
+    DATASET_TYPE_IDS_MAP = get_dataset_type_ids_map(folder)
+
     if gpc.get_global_rank() == 0:
         triples = [list(os.walk(folder, followlinks=True))]
     else:
@@ -400,7 +402,7 @@ def get_packed_dataset_without_short_length(
                         len(catch_ml_keys) < 2
                     ), f"The file name `{fp}` matched the following resample keys:{catch_ml_keys}"
 
-                ds_type_id = get_dataset_type_id(path=fp)
+                ds_type_id = get_dataset_type_id(DATASET_TYPE_IDS_MAP, path=fp)
                 ds = JsonlDataset(fp, ds_type_id, min_length=min_length_num)
 
                 if hasattr(ds, "old_length"):

internlm/data/utils.py

@@ -1,21 +1,26 @@
 #!/usr/bin/env python
 # -*- encoding: utf-8 -*-
+import os
+import re
 
 import torch
 
 from internlm.core.context import global_context as gpc
 
-DATASET_TYPE_IDS_MAP = {"en": 0, "cn": 1, "code": 2}
+
+def get_dataset_type_ids_map(path):
+    dirlist = list(os.listdir(path))
+    dirlist.sort()
+    return {key: idx for idx, key in enumerate(dirlist)}
 
 
-def get_dataset_type_id(path):
-    import re
-
+def get_dataset_type_id(dataset_type_ids_map, path):
     match_idxes = []
-    for key, idx in DATASET_TYPE_IDS_MAP.items():
+
+    for key, idx in dataset_type_ids_map.items():
         if re.search(rf"/[z_]*{key}/", path):
             match_idxes.append(idx)
-    assert len(match_idxes) == 1, f"{path}, match_idxes should be 1, but got {match_idxes} from {DATASET_TYPE_IDS_MAP}"
+    assert len(match_idxes) == 1, f"{path}, match_idxes should be 1, but got {match_idxes} from {dataset_type_ids_map}"
     return match_idxes[0]
 
 

internlm/train/training_internlm.py

@@ -33,7 +33,7 @@ from internlm.data.packed_dataset import (
     PackedDatasetWithoutCuSeqlen,
     get_packed_dataset_without_short_length,
 )
-from internlm.data.utils import DATASET_TYPE_IDS_MAP, unpack_data
+from internlm.data.utils import get_dataset_type_ids_map, unpack_data
 from internlm.model.embedding import Embedding1D
 from internlm.model.linear import (
     FeedForward,
@@ -133,7 +133,7 @@ def wrap_FSDP_model(model: Union[nn.Module, nn.ModuleList]):
 
         # wrap the model
         grp = gpc.get_group(ParallelMode.ZERO1)
-        model = FSDP(
+        model = FSDP(  # pylint: disable=unexpected-keyword-arg
             module=model,
             process_group=grp,
             sharding_strategy=ShardingStrategy.FULL_SHARD,
@@ -219,11 +219,11 @@ def get_train_data_loader(
 
     # Get the dataset types
     dataset_types = None
-    dataset_types = list(DATASET_TYPE_IDS_MAP.keys())
     data_cfg = gpc.config.data
 
     # Get the sample weight dictionary
     train_folder = data_cfg.train_folder
+    dataset_types = list(get_dataset_type_ids_map(train_folder).keys())
 
     if not train_folder:
         train_ds = RandomDataset(num_samples=1000000, max_len=data_cfg.seq_len)