Skip to content

Commit a1181cc

Browse files
binary-huskybinary-husky
committed
rename dir
1 parent 5883544 commit a1181cc

8 files changed

Lines changed: 10 additions & 10 deletions

File tree

‎ajet/copilot/write-swarm-client/SKILL.md‎

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -15,7 +15,7 @@ license: Complete terms in LICENSE.txt
1515
tutorial/user_math_agent
1616

1717
接下来，创建Agent源文件：
18-
tutorial/user_math_agent/agent_roll.py (以 tutorial/example_academic_trans/trans_roll.py 为模板，变化不大，关键是向用户索取必要的参数)
18+
tutorial/user_math_agent/agent_roll.py (以 tutorial/example_academic_trans_swarm/trans_roll.py 为模板，变化不大，关键是向用户索取必要的参数)
1919
tutorial/user_math_agent/agent_run.py （根据用户的要求，创建运行智能体的函数，或者类，都可以。同步异步都可以。）
2020
tutorial/user_math_agent/readme.md （Agent说明，以及训练、调试方法说明）
2121

@@ -37,7 +37,7 @@ def run_agent_and_compute_reward(task: Task, base_url:string, api_key:string) ->
3737
在 agent_roll 中，直接import run_agent_and_compute_reward即可。
3838

3939
- 智能体的编写要领：通过一个或几个Agent的协作，高效完成用户给定的任务。
40-
- 奖励编写的要领：容易验证的，使用规则直接计算。不容易验证的，模仿 `tutorial/example_academic_trans/train_multi_model/trans_reward.py` 中的方法，使用其他大型模型生成 LLM as Judge 程序。
40+
- 奖励编写的要领：容易验证的，使用规则直接计算。不容易验证的，模仿 `tutorial/example_academic_trans_swarm/train_multi_model/trans_reward.py` 中的方法，使用其他大型模型生成 LLM as Judge 程序。
4141

4242

4343
## 训练、调试方法说明
@@ -324,7 +324,7 @@ def run_agent_and_compute_reward(task: Task, base_url:string, api_key:string) ->
324324
from ajet.tuner_lib.experimental.as_swarm_client import SwarmClient, run_episodes_until_all_complete
325325
from ajet.default_config.ajet_default import AjetTaskReader, HuggingfaceDatRepo
326326
from ajet.task_reader import RouterTaskReader
327-
from tutorial.example_academic_trans.trans import execute_agent
327+
from tutorial.example_academic_trans_swarm.trans import execute_agent
328328

329329

330330
LOCAL_GRPO_N = 4 # grpo num_repeat (rollout.n)

‎docs/en/swarm_best_practice.md‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -135,7 +135,7 @@ from ajet.copilot.job import AgentJetJob
135135
from ajet.tuner_lib.experimental.as_swarm_client import SwarmClient, run_episodes_until_all_complete
136136
from ajet.default_config.ajet_default import AjetTaskReader, HuggingfaceDatRepo
137137
from ajet.task_reader import RouterTaskReader
138-
from tutorial.example_academic_trans.trans import execute_agent
138+
from tutorial.example_academic_trans_swarm.trans import execute_agent
139139

140140

141141
LOCAL_GRPO_N = 4 # grpo num_repeat (rollout.n)

tutorial/example_academic_trans/readme.md renamed to tutorial/example_academic_trans_swarm/readme.md

File renamed without changes.

tutorial/example_academic_trans/trans.py renamed to tutorial/example_academic_trans_swarm/trans.py

File renamed without changes.

tutorial/example_academic_trans/trans_reward.py renamed to tutorial/example_academic_trans_swarm/trans_reward.py

File renamed without changes.

tutorial/example_academic_trans/trans_roll.py renamed to tutorial/example_academic_trans_swarm/trans_roll.py

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -2,9 +2,9 @@
22
from ajet.tuner_lib.experimental.as_swarm_client import SwarmClient, run_episodes_until_all_complete
33
from ajet.default_config.ajet_default import AjetTaskReader, HuggingfaceDatRepo
44
from ajet.task_reader import RouterTaskReader
5-
from tutorial.example_academic_trans.trans import execute_agent
5+
from tutorial.example_academic_trans_swarm.trans import execute_agent
66

7-
# python -m tutorial.example_academic_trans.trans_roll
7+
# python -m tutorial.example_academic_trans_swarm.trans_roll
88

99

1010
# --------- configurations that take effect locally -------------

‎tutorial/example_train_multi_model/README.md‎

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -124,8 +124,8 @@ export DASHSCOPE_API_KEY="your-api-key-here"
124124
### 5. Run Training
125125

126126
```bash
127-
# tutorial/example_academic_trans/train-multi-model/trans_roll.py
128-
python -m tutorial.example_academic_trans.train_multi_model.trans_roll
127+
# tutorial/example_academic_trans_swarm/train-multi-model/trans_roll.py
128+
python -m tutorial.example_academic_trans_swarm.train_multi_model.trans_roll
129129
```
130130

131131
## How It Works

‎tutorial/example_train_multi_model/trans_roll.py‎

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -2,13 +2,13 @@
22
from ajet.tuner_lib.experimental.as_swarm_client import SwarmClient, run_episodes_until_all_complete
33
from ajet.default_config.ajet_default import AjetTaskReader, HuggingfaceDatRepo
44
from ajet.task_reader import RouterTaskReader
5-
from tutorial.example_academic_trans.train_multi_model.trans import execute_agent
5+
from tutorial.example_academic_trans_swarm.trans import execute_agent
66

77

88
# Step 1: Start two swarm servers on different ports:
99
# ajet-swarm start --swarm-port=10086 # For 7B model
1010
# ajet-swarm start --swarm-port=10086 # For 14B model
11-
# Step 2: python -m tutorial.example_academic_trans.train_multi_model.trans_roll
11+
# Step 2: python -m tutorial.example_academic_trans_swarm.train_multi_model.trans_roll
1212

1313
# --------- configurations that take effect locally -------------
1414
LOCAL_GRPO_N = 4 # grpo group size

0 commit comments

Comments
 (0)