Comments (5)
Hi @Sean-Reilly , thanks for the config! I am able to reproduce the bug. Looks like an issue when using n_cores_batch > 1
and the const
token. I can reproduce the bug with a simplified config:
{
"task" : {
"function_set" : ["add", "mul", "div", "sub", "const"]
},
"training" : {
"n_samples" : 100,
"batch_size" : 10,
"n_cores_batch" : 2
}
}
I will look into this and report back! Thanks.
from deep-symbolic-optimization.
Hi @Sean-Reilly, as a temporary hack, you can add pool = None
in train.py
at the beginning of the learn()
function. That will break some other use cases (namely, the control
task when using PyBullet envs), but should be just fine for regression
.
A real fix will be incoming.
from deep-symbolic-optimization.
Hi @Sean-Reilly , can you send me the full config you're using and the method to run (CLI vs Python interface)?
from deep-symbolic-optimization.
I tried in both CLI and Python interface and got the same error as above. Below is the full config.json
of a failed run.
{
"experiment": {
"logdir": "./log",
"seed": 0,
"timestamp": "2021-08-18-142149",
"task_name": "~_nas_rism-symbolic-regression_dso_data_hs_dens_0.1",
"save_path": "./log/~_nas_rism-symbolic-regression_dso_data_hs_dens_0.1_2021-08-18-142149"
},
"task": {
"task_type": "regression",
"dataset": "~/nas/rism-symbolic-regression/dso/data/hs_dens_0.1.csv",
"function_set": [
"add",
"mul",
"div",
"sub",
"const"
],
"metric": "inv_nrmse",
"metric_params": [
1.0
],
"extra_metric_test": null,
"extra_metric_test_params": [],
"threshold": 1e-20,
"protected": false,
"reward_noise": 0.0,
"reward_noise_type": "r",
"normalize_variance": false
},
"training": {
"n_epochs": null,
"n_samples": 100000,
"batch_size": 100,
"epsilon": 0.05,
"baseline": "R_e",
"alpha": 0.5,
"b_jumpstart": false,
"n_cores_batch": 4,
"complexity": "token",
"const_optimizer": "scipy",
"const_params": {},
"verbose": true,
"debug": 0,
"early_stopping": true,
"hof": 5,
"use_memory": false,
"memory_capacity": 1000.0,
"warm_start": null,
"memory_threshold": null,
"save_all_epoch": false,
"save_summary": false,
"save_positional_entropy": false,
"save_pareto_front": true,
"save_cache": false,
"save_cache_r_min": 0.9,
"save_freq": 1,
"runs": 100
},
"controller": {
"max_length": 30,
"cell": "lstm",
"num_layers": 1,
"num_units": 32,
"initializer": "zeros",
"embedding": false,
"embedding_size": 8,
"learning_rate": 0.0005,
"optimizer": "adam",
"observe_action": false,
"observe_parent": true,
"observe_sibling": true,
"entropy_weight": 0.03,
"entropy_gamma": 0.7,
"ppo": false,
"ppo_clip_ratio": 0.2,
"ppo_n_iters": 10,
"ppo_n_mb": 4,
"pqt": false,
"pqt_k": 10,
"pqt_batch_size": 1,
"pqt_weight": 200.0,
"pqt_use_pg": false,
"summary": false
},
"prior": {
"relational": {
"targets": [],
"effectors": [],
"relationship": null
},
"length": {
"min_": 4,
"max_": 30
},
"repeat": {
"tokens": "const",
"min_": null,
"max_": 3
},
"inverse": {},
"trig": {},
"const": {},
"no_inputs": {},
"uniform_arity": {},
"soft_length": {
"loc": 10,
"scale": 5
},
"language_model": {
"weight": null
}
},
"postprocess": {
"show_count": 5,
"save_plots": true
}
}
from deep-symbolic-optimization.
This should be fixed in the newest release.
from deep-symbolic-optimization.
Related Issues (20)
- # AttributeError: 'NoneType' object has no attribute 'execute' HOT 5
- Summing neg log p HOT 1
- How to set a time limit? HOT 1
- Kernel crashes when importing tensorflow version 1.14 HOT 5
- Running dso.run will create a log folder but does not store or load the results csv files HOT 7
- Deap fitness objective HOT 1
- create primitive set mapping name clash HOT 2
- Deap varying with constraints HOT 4
- Methods for Running DSO HOT 4
- Normalization for input variables to domain (0,1) HOT 1
- Normalization for input variables to domain (0,1) HOT 1
- Different Learned Equation with Different Numpy Array Shape HOT 4
- Doubts about output HOT 1
- Ratio of training set to testing set HOT 1
- training iteration N, current best R : 0.9 HOT 1
- Trouble Installing, sample virtual environment config? HOT 7
- How to install configuration and generate an interactive platform? HOT 1
- Defining custom gaussian function HOT 3
- Ignoring errors... HOT 1
- adjusting best R or custom best R HOT 1
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.
from deep-symbolic-optimization.