Full Experiment: Subband Allocation Results

Metrics and Training Values for the Subband Allocation Problem. This report only shows some models and runs that were preserved. Some of them were deleted due to various errors.

Rubén Cid

Created on February 10|Last edited on February 10

Comment

﻿
﻿
Mean Bit Rate in Validation
Mean Bit Rate in Validation
Showing first 5 runs
020406080Epoch20406080Bit Rate (Mbps)
name: p5-fnn-00-01-base
name: p4-set-00-01-base
name: p3-fnn-00-01-base
name: p1-joint-000-v1
name: p1-alloc-cnn-00-01-base
Results vs Ideal Conditions
Results vs Ideal Conditions
Showing first 5 runs
020406080Epoch11.522.533.5Proportional Loss Factor
name: p5-fnn-00-01-base
name: p4-set-00-01-base
name: p3-fnn-00-01-base
name: p1-joint-000-v1
name: p1-alloc-cnn-00-01-base
Mean Spectral Efficency in Validation
Mean Spectral Efficency in Validation
Showing first 5 runs
020406080Epoch2468Spectral Efficency
name: p5-fnn-00-01-base
name: p4-set-00-01-base
name: p3-fnn-00-01-base
name: p1-joint-000-v1
name: p1-alloc-cnn-00-01-base
Fairness of Allocations
Fairness of Allocations
Showing first 5 runs
020406080Epoch1234Jain Fairness
name: p5-fnn-00-01-base
name: p4-set-00-01-base
name: p3-fnn-00-01-base
name: p1-joint-000-v1
name: p1-alloc-cnn-00-01-base
Run set90
﻿
With the current results, we don't get better estimates than SISA Algorithm for the current setup. 
﻿

Add a comment