github/ColossalAI
1
0
Fork 0
mirror of https://github.com/hpcaitech/ColossalAI.git synced 2025-09-11 05:49:55 +00:00
Code Issues Packages Projects Releases Wiki Activity

[shardformer]update t5 tests for using all optimizations. (#4407)

Browse Source 
* [shardformer] gpt2 tests fix

[shardformer] test all optimizations (#4399)

[shardformer] test all optimizations

[shardformer] test all optimizations

[shardformer] test all optimizations

[shardformer] gpt2 tests fix

* [shardformer]update t5 to use all optimizations
This commit is contained in:
flybird11111
2023-08-14 15:49:13 +08:00
committed by Hongxin Liu
parent 1edc9b5fb3
commit 108e54a0b4
3 changed files with 33 additions and 16 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
tests/kit/model_zoo/transformers/t5.py
View File

@@ -16,8 +16,8 @@ def data_gen_for_encoder_only():
# config = T5Config(decoder_start_token_id=0)
# tokenizer = T5Tokenizer.from_pretrained("t5-small")
# input_ids = tokenizer("translate English to German: The house is wonderful.", return_tensors="pt").input_ids
input_ids = torch.Tensor([[13959, 1566, 12, 2968, 10, 37, 629, 19, 1627, 5, 1, 12]]).long()
attention_mask = torch.Tensor([[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0]]).long()
input_ids = torch.Tensor([[13959, 1566, 12, 2968, 10, 37, 629, 19, 1627, 5, 1, 12, 1627, 5, 1, 12]]).long()
attention_mask = torch.Tensor([[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0]]).long()
return dict(input_ids=input_ids, attention_mask=attention_mask)
@@ -26,7 +26,7 @@ def data_gen_for_conditional_generation():
#
# labels = tokenizer("Das Haus ist wunderbar.", return_tensors="pt").input_ids
data = data_gen_for_encoder_only()
labels = torch.Tensor([[644, 4598, 229, 19250, 5, 1, 644, 4598, 229, 19250, 5, 1]]).long()
labels = torch.Tensor([[644, 4598, 229, 19250, 5, 1, 644, 4598, 229, 19250, 5, 1, 229, 19250, 5, 1]]).long()
data['labels'] = labels
return data
@@ -35,7 +35,7 @@ def data_gen_for_t5_model():
# decoder_inputs_ids is obtained with the following code
# decoder_input_ids = model._shift_right(input_ids)
data = data_gen_for_encoder_only()
decoder_input_ids = torch.Tensor([[0, 13959, 1566, 12, 2968, 10, 37, 629, 19, 1627, 5, 5]]).long()
decoder_input_ids = torch.Tensor([[0, 13959, 1566, 12, 2968, 10, 37, 629, 19, 1627, 5, 5, 19, 1627, 5, 5]]).long()
data['decoder_input_ids'] = decoder_input_ids
return data