rwkv / lm-evaluation-harness Goto Github PK
View Code? Open in Web Editor NEWThis project forked from eleutherai/lm-evaluation-harness
A framework for few-shot evaluation of language models.
Home Page: https://www.eleuther.ai
License: MIT License
lm-evaluation-harness's People
Contributors
Stargazers
lm-evaluation-harness's Issues
Potential error with accelerate + lm-eval + rwkv sharded models
The following error occurs when running lm-eval with accelerate, with a sharded RWKV model
# ------------------------------
# Running Task : anli
# ------------------------------
The following values were not passed to `accelerate launch` and had defaults used instead:
`--num_processes` was set to a value of `7`
More than one GPU was found, enabling multi-GPU training.
If this was unintended please pass in `--num_processes=1`.
`--num_machines` was set to a value of `1`
`--mixed_precision` was set to a value of `'no'`
`--dynamo_backend` was set to a value of `'no'`
To avoid this warning pass in values for each of the problematic parameters or run `accelerate config`.
Using RTX 4000 series which doesn't support faster communication speedups. Ensuring P2P and IB communications are disabled.
2024-02-25:19:16:47,809 INFO [utils.py:145] Note: detected 160 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable.
2024-02-25:19:16:47,809 INFO [utils.py:148] Note: NumExpr detected 160 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8.
2024-02-25:19:16:47,887 INFO [utils.py:145] Note: detected 160 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable.
2024-02-25:19:16:47,887 INFO [utils.py:148] Note: NumExpr detected 160 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8.
2024-02-25:19:16:47,897 INFO [utils.py:145] Note: detected 160 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable.
2024-02-25:19:16:47,897 INFO [utils.py:148] Note: NumExpr detected 160 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8.
2024-02-25:19:16:47,904 INFO [utils.py:145] Note: detected 160 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable.
2024-02-25:19:16:47,905 INFO [utils.py:148] Note: NumExpr detected 160 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8.
2024-02-25:19:16:47,905 INFO [utils.py:145] Note: detected 160 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable.
2024-02-25:19:16:47,905 INFO [utils.py:148] Note: NumExpr detected 160 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8.
2024-02-25:19:16:47,906 INFO [utils.py:145] Note: detected 160 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable.
2024-02-25:19:16:47,906 INFO [utils.py:148] Note: NumExpr detected 160 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8.
2024-02-25:19:16:47,937 INFO [utils.py:145] Note: detected 160 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable.
2024-02-25:19:16:47,937 INFO [utils.py:148] Note: NumExpr detected 160 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8.
2024-02-25:19:16:48,003 INFO [config.py:58] PyTorch version 2.1.2 available.
2024-02-25:19:16:48,088 INFO [config.py:58] PyTorch version 2.1.2 available.
2024-02-25:19:16:48,097 INFO [config.py:58] PyTorch version 2.1.2 available.
2024-02-25:19:16:48,098 INFO [config.py:58] PyTorch version 2.1.2 available.
2024-02-25:19:16:48,101 INFO [config.py:58] PyTorch version 2.1.2 available.
2024-02-25:19:16:48,102 INFO [config.py:58] PyTorch version 2.1.2 available.
2024-02-25:19:16:48,110 INFO [config.py:58] PyTorch version 2.1.2 available.
2024-02-25:19:16:49,485 INFO [__main__.py:162] Verbosity set to INFO
2024-02-25:19:16:49,485 INFO [__init__.py:358] lm_eval.tasks.initialize_tasks() is deprecated and no longer necessary. It will be removed in v0.4.2 release. TaskManager will instead be used.
2024-02-25:19:16:49,558 INFO [__main__.py:162] Verbosity set to INFO
2024-02-25:19:16:49,559 INFO [__init__.py:358] lm_eval.tasks.initialize_tasks() is deprecated and no longer necessary. It will be removed in v0.4.2 release. TaskManager will instead be used.
2024-02-25:19:16:49,568 INFO [__main__.py:162] Verbosity set to INFO
2024-02-25:19:16:49,568 INFO [__init__.py:358] lm_eval.tasks.initialize_tasks() is deprecated and no longer necessary. It will be removed in v0.4.2 release. TaskManager will instead be used.
2024-02-25:19:16:49,589 INFO [__main__.py:162] Verbosity set to INFO
2024-02-25:19:16:49,589 INFO [__init__.py:358] lm_eval.tasks.initialize_tasks() is deprecated and no longer necessary. It will be removed in v0.4.2 release. TaskManager will instead be used.
2024-02-25:19:16:49,599 INFO [__main__.py:162] Verbosity set to INFO
2024-02-25:19:16:49,599 INFO [__init__.py:358] lm_eval.tasks.initialize_tasks() is deprecated and no longer necessary. It will be removed in v0.4.2 release. TaskManager will instead be used.
2024-02-25:19:16:49,607 INFO [__main__.py:162] Verbosity set to INFO
2024-02-25:19:16:49,607 INFO [__init__.py:358] lm_eval.tasks.initialize_tasks() is deprecated and no longer necessary. It will be removed in v0.4.2 release. TaskManager will instead be used.
2024-02-25:19:16:49,681 INFO [__main__.py:162] Verbosity set to INFO
2024-02-25:19:16:49,[681](https://github.com/RWKV/lm-evaluation-harness/actions/runs/8040203638/job/21958032153#step:3:688) INFO [__init__.py:358] lm_eval.tasks.initialize_tasks() is deprecated and no longer necessary. It will be removed in v0.4.2 release. TaskManager will instead be used.
2024-02-25:19:16:54,373 INFO [__main__.py:238] Selected Tasks: ['anli']
2024-02-25:19:16:54,373 INFO [__main__.py:239] Loading selected tasks...
2024-02-25:19:16:54,478 INFO [__main__.py:238] Selected Tasks: ['anli']
2024-02-25:19:16:54,478 INFO [__main__.py:239] Loading selected tasks...
2024-02-25:19:16:54,487 INFO [__main__.py:238] Selected Tasks: ['anli']
2024-02-25:19:16:54,487 INFO [__main__.py:239] Loading selected tasks...
2024-02-25:19:16:54,490 INFO [__main__.py:238] Selected Tasks: ['anli']
2024-02-25:19:16:54,490 INFO [__main__.py:239] Loading selected tasks...
2024-02-25:19:16:54,501 INFO [__main__.py:238] Selected Tasks: ['anli']
2024-02-25:19:16:54,501 INFO [__main__.py:239] Loading selected tasks...
2024-02-25:19:16:54,507 INFO [__main__.py:238] Selected Tasks: ['anli']
2024-02-25:19:16:54,507 INFO [__main__.py:239] Loading selected tasks...
exitcode : 1 (pid: 45813)
error_file: <N/A>
traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html
[2]:
time : 2024-02-25_19:16:59
host : 0c7aa106d18a
rank : 2 (local_rank: 2)
exitcode : 1 (pid: 45814)
error_file: <N/A>
traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html
[3]:
time : 2024-02-25_19:16:59
host : 0c7aa106d18a
rank : 3 (local_rank: 3)
exitcode : 1 (pid: 45815)
error_file: <N/A>
traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html
[4]:
time : 2024-02-25_19:16:59
host : 0c7aa106d18a
rank : 4 (local_rank: 4)
exitcode : 1 (pid: 45816)
error_file: <N/A>
traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html
[5]:
time : 2024-02-25_19:16:59
host : 0c7aa106d18a
rank : 5 (local_rank: 5)
exitcode : 1 (pid: 45817)
error_file: <N/A>
traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html
[6]:
time : 2024-02-25_19:16:59
host : 0c7aa106d18a
rank : 6 (local_rank: 6)
exitcode : 1 (pid: 45818)
error_file: <N/A>
traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html
------------------------------------------------------------
Root Cause (first observed failure):
[0]:
time : 2024-02-25_19:16:59
host : 0c7aa106d18a
rank : 0 (local_rank: 0)
exitcode : 1 (pid: 45812)
error_file: <N/A>
traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html
============================================================
This error does not seem to occur with smaller unsharded models.
As such as a temporary work around, we would be avoiding sharding the checkpoint for the HF implementation.
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.