NousResearch/atropos
1
0
Fork 0
mirror of https://github.com/NousResearch/atropos.git synced 2026-04-19 12:57:58 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

change eval set size since this is a small dataset we need mo data for trainn

Browse source
This commit is contained in:
teknium1 2025-05-14 19:18:01 -07:00
parent bcc38567ca
commit 8a0e107806
1 changed files with 3 additions and 11 deletions
Download patch file Download diff file Expand all files Collapse all files

14
environments/instruction_following_algorithm_environment.py
View file

@ -59,6 +59,7 @@ class InstructionFollowingEnv(BaseEnv):
eval_limit_ratio=0.1,
dataset_name="allenai/RLVR-IFeval", # Default dataset
dataset_config_name=None, # RLVR-IFeval doesn't have a specific config name, uses 'default'
test_set_ratio=0.05 # The ratio of the selelcted dataset in %
)
# Server configurations can be similar to SingleToolCallingEnv or adjusted
server_configs = [
@ -68,14 +69,7 @@ class InstructionFollowingEnv(BaseEnv):
api_key="x",
num_max_requests_at_once=32,
num_requests_for_eval=256,
),
APIServerConfig(
model_name="NousResearch/DeepHermes-3-Llama-3-8B-Preview",
base_url="http://localhost:9005/v1",
api_key="x",
num_max_requests_at_once=32,
num_requests_for_eval=256,
),
)
]
return env_config, server_configs
@ -214,7 +208,7 @@ class InstructionFollowingEnv(BaseEnv):
full_dataset = full_dataset.shuffle(seed=42)
actual_test_size = 0.2
actual_test_size = self.config.test_set_ratio # Read from config
num_items = len(full_dataset)
if num_items == 0:
@ -813,8 +807,6 @@ def validate_quotation(text: str) -> bool:
def validate_no_commas(text: str) -> bool:
return "," not in text
IF_FUNCTIONS_MAP = {
"verify_keywords": verify_keywords,
"verify_keyword_frequency": verify_keyword_frequency,