Crystalcareai
commited on
Update generate.py
Browse files- generate.py +0 -9
generate.py
CHANGED
@@ -1,10 +1,4 @@
|
|
1 |
import torch
|
2 |
-
from transformers.generation.utils import (
|
3 |
-
GenerationMixin,
|
4 |
-
validate_stopping_criteria,
|
5 |
-
StoppingCriteriaList,
|
6 |
-
)
|
7 |
-
from transformers import TextStreamer
|
8 |
|
9 |
|
10 |
def custom_generate(
|
@@ -50,7 +44,6 @@ def custom_generate(
|
|
50 |
if max_new_tokens is None:
|
51 |
max_new_tokens = 50 # Default value if not specified
|
52 |
for cur_token_idx in range(max_new_tokens):
|
53 |
-
# Sample the next token
|
54 |
new_ids = self(
|
55 |
input_ids[~finished_generating],
|
56 |
attention_mask=attention_mask[~finished_generating] if attention_mask is not None else None,
|
@@ -95,14 +88,12 @@ def custom_generate(
|
|
95 |
if streamer is not None:
|
96 |
streamer.put(new_ids_sampled)
|
97 |
|
98 |
-
# Create a named tuple to match the expected output format
|
99 |
from collections import namedtuple
|
100 |
GenerateOutput = namedtuple("GenerateOutput", ["sequences", "scores", "attentions", "hidden_states"])
|
101 |
|
102 |
# Convert the generated token IDs to a tensor
|
103 |
generated_token_ids_tensor = input_ids
|
104 |
|
105 |
-
# Create the GenerateOutput named tuple
|
106 |
output = GenerateOutput(
|
107 |
sequences=generated_token_ids_tensor,
|
108 |
scores=None,
|
|
|
1 |
import torch
|
|
|
|
|
|
|
|
|
|
|
|
|
2 |
|
3 |
|
4 |
def custom_generate(
|
|
|
44 |
if max_new_tokens is None:
|
45 |
max_new_tokens = 50 # Default value if not specified
|
46 |
for cur_token_idx in range(max_new_tokens):
|
|
|
47 |
new_ids = self(
|
48 |
input_ids[~finished_generating],
|
49 |
attention_mask=attention_mask[~finished_generating] if attention_mask is not None else None,
|
|
|
88 |
if streamer is not None:
|
89 |
streamer.put(new_ids_sampled)
|
90 |
|
|
|
91 |
from collections import namedtuple
|
92 |
GenerateOutput = namedtuple("GenerateOutput", ["sequences", "scores", "attentions", "hidden_states"])
|
93 |
|
94 |
# Convert the generated token IDs to a tensor
|
95 |
generated_token_ids_tensor = input_ids
|
96 |
|
|
|
97 |
output = GenerateOutput(
|
98 |
sequences=generated_token_ids_tensor,
|
99 |
scores=None,
|