mirror of
https://github.com/bigscience-workshop/petals
synced 2024-11-16 06:12:50 +00:00
f0c7383181
- finish renaming RemoteSequenceInfo -> RemoteSequenceManager (why: if it was an *Info, user would expect it to be similar - to a dataclass; whereas in actuality, the class is doing heavy network interactions on its own) - implement RemoteSequenceManager.make_sequence (from https://pastebin.com/uXgy2U8B ) - make RemoteSequentialInferenceSession use RemoteSequenceManager.make_sequence - make tests pass again - make it possible to create inference session without RemoteTransformerBlock - make a standalone test for RemoteSequential - rollback convert-model Co-authored-by: Tim Dettmers <tim.dettmers@gmail.com>
14 lines
419 B
Python
14 lines
419 B
Python
import os
|
|
|
|
INITIAL_PEERS = os.environ.get("INITIAL_PEERS")
|
|
if not INITIAL_PEERS:
|
|
raise RuntimeError("Must specify INITIAL_PEERS environment variable with one or more peer ids")
|
|
INITIAL_PEERS = INITIAL_PEERS.split()
|
|
|
|
|
|
MODEL_NAME = os.environ.get("MODEL_NAME")
|
|
if not MODEL_NAME:
|
|
raise RuntimeError("Must specify MODEL_NAME as an index of a transformer block to be tested")
|
|
|
|
REF_NAME = os.environ.get("REF_NAME")
|