Skip to content

updated mlflow in miner py #23

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 2 commits into
base: ml_flow_integration
Choose a base branch
from
Open

updated mlflow in miner py #23

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
cbecf10
Updated ml_flow_integration
HeliosPrimeOne Mar 30, 2024
a15a52c
Update mlflow miner.py
HeliosPrimeOne Mar 30, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 10 additions & 4 deletions neurons/miner.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -41,7 +41,13 @@

args = Configurator.combine_configs()

mlflow.set_tracking_uri(uri=args.mlflow_server)
mlflow.set_tracking_uri(uri=args.mlflow_server) # ```mlflow ui``` defaults to port 5000
mlflow.start_run(run_name="HiveMind Trainer") # Names the Training instead of random name.

# Here's the MLflow dataset grabber
dataset_name = "tiiuae/falcon-refinedweb" # defines dataset name in parameters
mlflow.log_param("dataset", dataset_name) # defines dataset name in parameters


def flatten_list(nested_list):
"""Flatten a nested list."""
Expand All @@ -55,15 +61,15 @@ def flatten_list(nested_list):
inital_peers_request = requests.get(args.miner.bootstrapping_server)
initial_peers = inital_peers_request.json()["initial_peers"]
assert not (initial_peers is None)
#initial_peers = flatten_list(args.initial_peers)
# initial_peers = flatten_list(args.initial_peers)
batch_size = args.batch_size
save_every = args.save_every
block_size = 512
num_steps = 100_000_000_000 #infinite training
target_batch_size = 81920 #when to average all weights.

dataset_config = {
"dataset": "tiiuae/falcon-refinedweb",
"dataset": dataset_name,
"key": "content",
"split": "train",
"block_size": block_size,
Expand Down Expand Up @@ -479,7 +485,7 @@ def on_train_batch_end(self, trainer, lm, outputs, batch, batch_idx, checksum=No

self.last_report_time = current_time

mlflow.log_metric("loss",outputs["loss"].item(), step=self.step )
mlflow.log_metric("loss",outputs["loss"].item(), step=self.step ) # Logs the loss metrics from training <--step=batch- not global step.
for url in self.validator_urls:
try:
response = requests.post(
Expand Down