github/ColossalAI
1
0
Fork 0
mirror of https://github.com/hpcaitech/ColossalAI.git synced 2026-07-16 17:16:14 +00:00
Code Issues Packages Projects Releases Wiki Activity

[pre-commit.ci] auto fixes from pre-commit.com hooks

Browse Source 
for more information, see https://pre-commit.ci
This commit is contained in:
pre-commit-ci[bot]
2026-07-13 18:00:14 +00:00
parent 90dacce9c2
commit 9009ab6cfb
50 changed files with 188 additions and 214 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
applications/Colossal-LLaMA/colossal_llama/model/init_model.py
View File

@@ -4,6 +4,7 @@
"""
Initialize new model with updated tokenizer by calculating the mean values from original model
"""
import argparse
import numpy as np

28
applications/ColossalChat/coati/dataset/tokenization_utils.py
View File

@@ -56,10 +56,8 @@ def tokenize_sft(
template.messages = []
for idx, mess in enumerate(messages):
if mess["from"] != template.roles[idx % 2]:
raise ValueError(
f"Message should iterate between user and assistant and starts with a \
line from the user. Got the following data:\n{messages}"
)
raise ValueError(f"Message should iterate between user and assistant and starts with a \
line from the user. Got the following data:\n{messages}")
template.append_message(mess["from"], mess["content"])
if len(template.messages) % 2 != 0:
@@ -245,10 +243,8 @@ def tokenize_rlhf(
for idx, mess in enumerate(context):
if mess["from"] != template.roles[idx % 2]:
raise ValueError(
f"Message should iterate between user and assistant and starts with a \
line from the user. Got the following data:\n{context}"
)
raise ValueError(f"Message should iterate between user and assistant and starts with a \
line from the user. Got the following data:\n{context}")
template.append_message(mess["from"], mess["content"])
if len(template.messages) % 2 != 1:
@@ -272,18 +268,14 @@ def tokenize_rlhf(
rejected_continuation = data_point["rejected"]
for round in range(len(chosen_continuation)):
if chosen_continuation[round]["from"] != template.roles[(round + 1) % 2]:
raise ValueError(
f"Message should iterate between user and assistant and starts with a \
line from the user. Got the following data:\n{chosen_continuation}"
)
raise ValueError(f"Message should iterate between user and assistant and starts with a \
line from the user. Got the following data:\n{chosen_continuation}")
chosen.append_message(chosen_continuation[round]["from"], chosen_continuation[round]["content"])
for round in range(len(rejected_continuation)):
if rejected_continuation[round]["from"] != template.roles[(round + 1) % 2]:
raise ValueError(
f"Message should iterate between user and assistant and starts with a \
line from the user. Got the following data:\n{rejected_continuation}"
)
raise ValueError(f"Message should iterate between user and assistant and starts with a \
line from the user. Got the following data:\n{rejected_continuation}")
rejected.append_message(rejected_continuation[round]["from"], rejected_continuation[round]["content"])
(
@@ -296,14 +288,14 @@ def tokenize_rlhf(
) = (None, None, None, None, None, None)
chosen_data_packed = apply_rlhf_data_format(chosen, tokenizer)
(chosen_input_ids, chosen_loss_mask, chosen_label_decode) = (
chosen_input_ids, chosen_loss_mask, chosen_label_decode = (
chosen_data_packed["input_ids"],
chosen_data_packed["loss_mask"],
chosen_data_packed["label_decode"],
)
rejected_data_packed = apply_rlhf_data_format(rejected, tokenizer)
(rejected_input_ids, rejected_loss_mask, rejected_label_decode) = (
rejected_input_ids, rejected_loss_mask, rejected_label_decode = (
rejected_data_packed["input_ids"],
rejected_data_packed["loss_mask"],
rejected_data_packed["label_decode"],

1
applications/ColossalChat/coati/distributed/reward/reward_fn.py
View File

@@ -17,7 +17,6 @@ under the Apache License 2.0:
https://github.com/volcengine/verl
"""
import json
import torch

4
applications/ColossalChat/coati/trainer/kto.py
View File

@@ -130,7 +130,7 @@ class KTOTrainer(SLTrainer):
)
for i, batch in enumerate(self.train_dataloader):
batch = to_device(batch, self.device)
(input_ids, attention_mask, loss_mask, label, kl_input_ids, kl_attention_mask, kl_loss_mask) = (
input_ids, attention_mask, loss_mask, label, kl_input_ids, kl_attention_mask, kl_loss_mask = (
batch["input_ids"],
batch["attention_mask"],
batch["loss_mask"],
@@ -279,7 +279,7 @@ class KTOTrainer(SLTrainer):
)
for i, batch in enumerate(self.train_dataloader):
batch = to_device(batch, self.device)
(input_ids, attention_mask, loss_mask, label, kl_input_ids, kl_attention_mask, kl_loss_mask) = (
input_ids, attention_mask, loss_mask, label, kl_input_ids, kl_attention_mask, kl_loss_mask = (
batch["input_ids"],
batch["attention_mask"],
batch["loss_mask"],

2
applications/ColossalChat/examples/community/ray/train_prompts_on_ray.py
View File

@@ -120,7 +120,7 @@ class BasePPORole(DistributedTorchRayActor):
def _prepare_model_with_strategy(self, has_optimizer: bool):
if has_optimizer:
self._init_optimizer()
(self._model, self._optimizer) = self._strategy.prepare((self._model, self._optimizer))
self._model, self._optimizer = self._strategy.prepare((self._model, self._optimizer))
else:
self._model = self._strategy.prepare(self._model)

4
applications/ColossalQA/examples/webui_demo/webui.py
View File

@@ -81,11 +81,11 @@ with gr.Blocks(css=CSS) as demo:
)
with gr.Row():
btn = gr.UploadButton("📁", file_types=["file"], file_count="multiple", size="sm")
restart_btn = gr.Button(str("\u21BB"), elem_id="restart-btn", scale=1)
restart_btn = gr.Button(str("\u21bb"), elem_id="restart-btn", scale=1)
txt = gr.Textbox(
scale=8,
show_label=False,
placeholder="Enter text and press enter, or use 📁 to upload files, click \u21BB to clear loaded files and restart chat",
placeholder="Enter text and press enter, or use 📁 to upload files, click \u21bb to clear loaded files and restart chat",
container=True,
autofocus=True,
)