dataset stringclasses 3
values | method_group stringclasses 2
values | variant_name stringlengths 8 36 | model_name stringclasses 1
value | batch_size int64 64 128 | amp_enabled bool 2
classes | epochs_run int64 50 50 | best_top1 float64 0.24 0.9 | best_top5 float64 0.51 1 | final_f1 float64 0.18 0.9 | total_energy_kwh float64 0.1 0.41 | total_co2_kg float64 0.05 0.2 | total_time_sec float64 4.58k 19.8k | peak_vram_mb float64 3.27k 4.44k | status stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
cifar10 | cumulative_ablation | C0_baseline | effnetv2_s | 64 | false | 50 | 0.7569 | 0.9865 | 0.756763 | 0.198818 | 0.094439 | 9,591.358462 | 4,275.367188 | completed |
cifar10 | cumulative_ablation | C1_cache | effnetv2_s | 64 | false | 50 | 0.7472 | 0.9862 | 0.715194 | 0.198469 | 0.094273 | 9,556.001816 | 4,356.42627 | completed |
cifar10 | cumulative_ablation | C2_cache_amp | effnetv2_s | 128 | true | 50 | 0.8068 | 0.9913 | 0.780503 | 0.098796 | 0.046928 | 4,823.602003 | 4,141.093262 | completed |
cifar10 | cumulative_ablation | C3_cache_amp_gradaccum | effnetv2_s | 128 | true | 50 | 0.8345 | 0.9937 | 0.82655 | 0.098766 | 0.046914 | 4,810.303344 | 3,451.180176 | completed |
cifar10 | cumulative_ablation | C4_cache_amp_gradaccum_adaptivelr | effnetv2_s | 128 | true | 50 | 0.8973 | 0.9971 | 0.896878 | 0.098928 | 0.046991 | 4,825.418664 | 3,451.180176 | completed |
cifar10 | cumulative_ablation | C5_cache_amp_gradaccum_adaptivelr_l1 | effnetv2_s | 128 | true | 50 | 0.904 | 0.9977 | 0.90076 | 0.10062 | 0.047794 | 4,909.434878 | 3,517.291504 | completed |
cifar10 | cumulative_ablation | C6_full_e2am | effnetv2_s | 128 | true | 50 | 0.904 | 0.9977 | 0.90076 | 0.100711 | 0.047838 | 4,914.646834 | 3,517.291504 | completed |
cifar10 | individual_methods | M0_baseline_fp32 | effnetv2_s | 64 | false | 50 | 0.7565 | 0.9839 | 0.710483 | 0.20802 | 0.09881 | 9,714.07477 | 3,274.149414 | completed |
cifar10 | individual_methods | M1_cache_only | effnetv2_s | 64 | false | 50 | 0.7582 | 0.9851 | 0.721166 | 0.19628 | 0.093233 | 9,440.229767 | 3,279.649414 | completed |
cifar10 | individual_methods | M2_amp_only | effnetv2_s | 128 | true | 50 | 0.8013 | 0.9909 | 0.782524 | 0.098642 | 0.046855 | 4,774.621377 | 4,141.093262 | completed |
cifar10 | individual_methods | M3_grad_accum_only | effnetv2_s | 64 | false | 50 | 0.8036 | 0.9903 | 0.80465 | 0.195459 | 0.092843 | 9,402.74931 | 3,377.65918 | completed |
cifar10 | individual_methods | M4_l1_sparsity_only | effnetv2_s | 64 | false | 50 | 0.7394 | 0.9837 | 0.711569 | 0.209743 | 0.099628 | 9,834.075927 | 4,436.501465 | completed |
cifar10 | individual_methods | M5_adaptive_lr_only | effnetv2_s | 64 | false | 50 | 0.8841 | 0.9969 | 0.883535 | 0.196328 | 0.093256 | 9,453.784048 | 4,275.367188 | completed |
cifar10 | individual_methods | M6_eag_only | effnetv2_s | 64 | false | 50 | 0.759 | 0.9854 | 0.748992 | 0.20804 | 0.098819 | 9,713.186405 | 3,274.149414 | completed |
cifar10 | individual_methods | M7_full_e2am | effnetv2_s | 128 | true | 50 | 0.8974 | 0.9967 | 0.897066 | 0.100517 | 0.047745 | 4,846.892284 | 3,517.291504 | completed |
cifar100 | cumulative_ablation | C0_baseline | effnetv2_s | 64 | false | 50 | 0.4075 | 0.7359 | 0.380174 | 0.199915 | 0.09496 | 9,529.088878 | 4,276.268066 | completed |
cifar100 | cumulative_ablation | C1_cache | effnetv2_s | 64 | false | 50 | 0.4502 | 0.77 | 0.422574 | 0.199817 | 0.094913 | 9,533.118624 | 3,280.489746 | completed |
cifar100 | cumulative_ablation | C2_cache_amp | effnetv2_s | 128 | true | 50 | 0.5411 | 0.8401 | 0.53222 | 0.098153 | 0.046623 | 4,755.724461 | 4,220.992676 | completed |
cifar100 | cumulative_ablation | C3_cache_amp_gradaccum | effnetv2_s | 128 | true | 50 | 0.6111 | 0.8775 | 0.607742 | 0.095349 | 0.045291 | 4,577.647082 | 3,447.270508 | completed |
cifar100 | cumulative_ablation | C4_cache_amp_gradaccum_adaptivelr | effnetv2_s | 128 | true | 50 | 0.7096 | 0.9251 | 0.709578 | 0.095406 | 0.045318 | 4,587.17475 | 3,447.270508 | completed |
cifar100 | cumulative_ablation | C5_cache_amp_gradaccum_adaptivelr_l1 | effnetv2_s | 128 | true | 50 | 0.7016 | 0.9205 | 0.700857 | 0.096829 | 0.045994 | 4,671.625259 | 3,514.631836 | completed |
cifar100 | cumulative_ablation | C6_full_e2am | effnetv2_s | 128 | true | 50 | 0.7017 | 0.9191 | 0.700828 | 0.099904 | 0.047454 | 4,811.688165 | 4,303.745605 | completed |
cifar100 | individual_methods | M0_baseline_fp32 | effnetv2_s | 64 | false | 50 | 0.39 | 0.7104 | 0.348463 | 0.198957 | 0.094505 | 9,331.550753 | 4,357.766602 | completed |
cifar100 | individual_methods | M1_cache_only | effnetv2_s | 64 | false | 50 | 0.419 | 0.7465 | 0.390691 | 0.199188 | 0.094614 | 9,688.777062 | 3,275.489746 | completed |
cifar100 | individual_methods | M2_amp_only | effnetv2_s | 128 | true | 50 | 0.524 | 0.8238 | 0.510378 | 0.095633 | 0.045426 | 4,607.844692 | 4,141.994141 | completed |
cifar100 | individual_methods | M3_grad_accum_only | effnetv2_s | 64 | false | 50 | 0.5408 | 0.8365 | 0.529794 | 0.198441 | 0.09426 | 9,629.036025 | 4,357.766602 | completed |
cifar100 | individual_methods | M4_l1_sparsity_only | effnetv2_s | 64 | false | 50 | 0.4378 | 0.7671 | 0.406935 | 0.201874 | 0.09589 | 9,629.383928 | 4,437.841797 | completed |
cifar100 | individual_methods | M5_adaptive_lr_only | effnetv2_s | 64 | false | 50 | 0.5975 | 0.8641 | 0.593194 | 0.199514 | 0.094769 | 9,677.777307 | 4,357.766602 | completed |
cifar100 | individual_methods | M6_eag_only | effnetv2_s | 64 | false | 50 | 0.401 | 0.7281 | 0.353792 | 0.199925 | 0.094964 | 9,519.850924 | 4,357.766602 | completed |
cifar100 | individual_methods | M7_full_e2am | effnetv2_s | 128 | true | 50 | 0.709 | 0.9254 | 0.709078 | 0.10245 | 0.048664 | 4,969.657219 | 3,511.881836 | completed |
tiny_imagenet | cumulative_ablation | C0_baseline | effnetv2_s | 64 | false | 50 | 0.2475 | 0.5199 | 0.211268 | 0.411919 | 0.195662 | 19,800.469194 | 4,358.757324 | completed |
tiny_imagenet | cumulative_ablation | C1_cache | effnetv2_s | 64 | false | 50 | 0.2502 | 0.5186 | 0.19491 | 0.401428 | 0.190678 | 19,311.573588 | 4,358.757324 | completed |
tiny_imagenet | cumulative_ablation | C2_cache_amp | effnetv2_s | 128 | true | 50 | 0.3465 | 0.63 | 0.318729 | 0.197953 | 0.094028 | 9,632.562859 | 4,146.996094 | completed |
tiny_imagenet | cumulative_ablation | C3_cache_amp_gradaccum | effnetv2_s | 128 | true | 50 | 0.4436 | 0.7242 | 0.439476 | 0.201201 | 0.09557 | 9,765.766818 | 3,453.76123 | completed |
tiny_imagenet | cumulative_ablation | C4_cache_amp_gradaccum_adaptivelr | effnetv2_s | 128 | true | 50 | 0.5738 | 0.8162 | 0.568648 | 0.202648 | 0.096258 | 9,831.056419 | 4,228.983398 | completed |
tiny_imagenet | cumulative_ablation | C5_cache_amp_gradaccum_adaptivelr_l1 | effnetv2_s | 128 | true | 50 | 0.5708 | 0.8125 | 0.565335 | 0.205338 | 0.097536 | 9,943.09728 | 3,518.372559 | completed |
tiny_imagenet | cumulative_ablation | C6_full_e2am | effnetv2_s | 128 | true | 50 | 0.5708 | 0.8125 | 0.565335 | 0.205262 | 0.0975 | 9,939.098759 | 3,518.372559 | completed |
tiny_imagenet | individual_methods | M0_baseline_fp32 | effnetv2_s | 64 | false | 50 | 0.2462 | 0.517 | 0.223854 | 0.395184 | 0.187712 | 18,948.885623 | 4,358.757324 | completed |
tiny_imagenet | individual_methods | M1_cache_only | effnetv2_s | 64 | false | 50 | 0.2451 | 0.5128 | 0.182945 | 0.394418 | 0.187348 | 19,007.64749 | 4,358.757324 | completed |
tiny_imagenet | individual_methods | M2_amp_only | effnetv2_s | 128 | true | 50 | 0.3541 | 0.6305 | 0.294041 | 0.193673 | 0.091995 | 9,387.331252 | 4,229.983398 | completed |
tiny_imagenet | individual_methods | M3_grad_accum_only | effnetv2_s | 64 | false | 50 | 0.3467 | 0.6401 | 0.324298 | 0.392124 | 0.186259 | 18,791.001306 | 3,383.490234 | completed |
tiny_imagenet | individual_methods | M4_l1_sparsity_only | effnetv2_s | 64 | false | 50 | 0.244 | 0.5172 | 0.210349 | 0.405434 | 0.192581 | 19,502.241866 | 4,440.33252 | completed |
tiny_imagenet | individual_methods | M5_adaptive_lr_only | effnetv2_s | 64 | false | 50 | 0.4611 | 0.7419 | 0.449641 | 0.397918 | 0.189011 | 19,208.212265 | 4,358.757324 | completed |
tiny_imagenet | individual_methods | M6_eag_only | effnetv2_s | 64 | false | 50 | 0.2606 | 0.5416 | 0.209208 | 0.39925 | 0.189644 | 19,318.716631 | 3,285.480469 | completed |
tiny_imagenet | individual_methods | M7_full_e2am | effnetv2_s | 128 | true | 50 | 0.5805 | 0.8162 | 0.575147 | 0.200459 | 0.095218 | 9,682.665335 | 3,519.122559 | completed |
E2AM Ablation Results: EfficientNetV2-S
Energy-aware training ablation study for EfficientNetV2-S across three image-classification datasets: CIFAR-10, CIFAR-100, and Tiny-ImageNet.
Each dataset has 15 training variants (8 individual-method M0..M7, 7 cumulative ablation C0..C6) at 50 epochs, plus a 5-variant deployment pipeline (FP32 baseline, structured pruning, pruning+finetune, INT8 quantization, pruned+INT8).
Status: 45 completed variants, 0 partial.
Quick links
- Methodology
- Headline results
- Cross-dataset comparison
- Per-dataset results
- Deployment results
- Reproducibility
Headline results
| Dataset | Best variant | Top-1 | Top-5 | Energy (kWh) | CO₂ (kg) | Time (sec) |
|---|---|---|---|---|---|---|
| CIFAR-10 | C5_cache_amp_gradaccum_adaptivelr_l1 | 0.9040 | 0.9977 | 0.1006 | 0.0478 | 4909 |
| CIFAR-100 | C4_cache_amp_gradaccum_adaptivelr | 0.7096 | 0.9251 | 0.0954 | 0.0453 | 4587 |
| Tiny-ImageNet | M7_full_e2am | 0.5805 | 0.8162 | 0.2005 | 0.0952 | 9683 |
Cross-dataset comparison
How the same training variants behave across CIFAR-10, CIFAR-100, and Tiny-ImageNet.
Accuracy By Variant Across Datasets
Energy By Variant Across Datasets
Per-dataset results
CIFAR-10
M-matrix (individual methods)
| Variant | Epochs | Top-1 | Top-5 | Energy (kWh) | CO₂ (kg) | Time (s) | Status |
|---|---|---|---|---|---|---|---|
| M0_baseline_fp32 | 50 | 0.7565 | 0.9839 | 0.2080 | 0.0988 | 9714 | completed |
| M1_cache_only | 50 | 0.7582 | 0.9851 | 0.1963 | 0.0932 | 9440 | completed |
| M2_amp_only | 50 | 0.8013 | 0.9909 | 0.0986 | 0.0469 | 4775 | completed |
| M3_grad_accum_only | 50 | 0.8036 | 0.9903 | 0.1955 | 0.0928 | 9403 | completed |
| M4_l1_sparsity_only | 50 | 0.7394 | 0.9837 | 0.2097 | 0.0996 | 9834 | completed |
| M5_adaptive_lr_only | 50 | 0.8841 | 0.9969 | 0.1963 | 0.0933 | 9454 | completed |
| M6_eag_only | 50 | 0.7590 | 0.9854 | 0.2080 | 0.0988 | 9713 | completed |
| M7_full_e2am | 50 | 0.8974 | 0.9967 | 0.1005 | 0.0477 | 4847 | completed |
C-matrix (cumulative ablation)
| Variant | Epochs | Top-1 | Top-5 | Energy (kWh) | CO₂ (kg) | Time (s) | Status |
|---|---|---|---|---|---|---|---|
| C0_baseline | 50 | 0.7569 | 0.9865 | 0.1988 | 0.0944 | 9591 | completed |
| C1_cache | 50 | 0.7472 | 0.9862 | 0.1985 | 0.0943 | 9556 | completed |
| C2_cache_amp | 50 | 0.8068 | 0.9913 | 0.0988 | 0.0469 | 4824 | completed |
| C3_cache_amp_gradaccum | 50 | 0.8345 | 0.9937 | 0.0988 | 0.0469 | 4810 | completed |
| C4_cache_amp_gradaccum_adaptivelr | 50 | 0.8973 | 0.9971 | 0.0989 | 0.0470 | 4825 | completed |
| C5_cache_amp_gradaccum_adaptivelr_l1 | 50 | 0.9040 | 0.9977 | 0.1006 | 0.0478 | 4909 | completed |
| C6_full_e2am | 50 | 0.9040 | 0.9977 | 0.1007 | 0.0478 | 4915 | completed |
CIFAR-100
M-matrix (individual methods)
| Variant | Epochs | Top-1 | Top-5 | Energy (kWh) | CO₂ (kg) | Time (s) | Status |
|---|---|---|---|---|---|---|---|
| M0_baseline_fp32 | 50 | 0.3900 | 0.7104 | 0.1990 | 0.0945 | 9332 | completed |
| M1_cache_only | 50 | 0.4190 | 0.7465 | 0.1992 | 0.0946 | 9689 | completed |
| M2_amp_only | 50 | 0.5240 | 0.8238 | 0.0956 | 0.0454 | 4608 | completed |
| M3_grad_accum_only | 50 | 0.5408 | 0.8365 | 0.1984 | 0.0943 | 9629 | completed |
| M4_l1_sparsity_only | 50 | 0.4378 | 0.7671 | 0.2019 | 0.0959 | 9629 | completed |
| M5_adaptive_lr_only | 50 | 0.5975 | 0.8641 | 0.1995 | 0.0948 | 9678 | completed |
| M6_eag_only | 50 | 0.4010 | 0.7281 | 0.1999 | 0.0950 | 9520 | completed |
| M7_full_e2am | 50 | 0.7090 | 0.9254 | 0.1025 | 0.0487 | 4970 | completed |
C-matrix (cumulative ablation)
| Variant | Epochs | Top-1 | Top-5 | Energy (kWh) | CO₂ (kg) | Time (s) | Status |
|---|---|---|---|---|---|---|---|
| C0_baseline | 50 | 0.4075 | 0.7359 | 0.1999 | 0.0950 | 9529 | completed |
| C1_cache | 50 | 0.4502 | 0.7700 | 0.1998 | 0.0949 | 9533 | completed |
| C2_cache_amp | 50 | 0.5411 | 0.8401 | 0.0982 | 0.0466 | 4756 | completed |
| C3_cache_amp_gradaccum | 50 | 0.6111 | 0.8775 | 0.0953 | 0.0453 | 4578 | completed |
| C4_cache_amp_gradaccum_adaptivelr | 50 | 0.7096 | 0.9251 | 0.0954 | 0.0453 | 4587 | completed |
| C5_cache_amp_gradaccum_adaptivelr_l1 | 50 | 0.7016 | 0.9205 | 0.0968 | 0.0460 | 4672 | completed |
| C6_full_e2am | 50 | 0.7017 | 0.9191 | 0.0999 | 0.0475 | 4812 | completed |
Tiny-ImageNet
M-matrix (individual methods)
| Variant | Epochs | Top-1 | Top-5 | Energy (kWh) | CO₂ (kg) | Time (s) | Status |
|---|---|---|---|---|---|---|---|
| M0_baseline_fp32 | 50 | 0.2462 | 0.5170 | 0.3952 | 0.1877 | 18949 | completed |
| M1_cache_only | 50 | 0.2451 | 0.5128 | 0.3944 | 0.1873 | 19008 | completed |
| M2_amp_only | 50 | 0.3541 | 0.6305 | 0.1937 | 0.0920 | 9387 | completed |
| M3_grad_accum_only | 50 | 0.3467 | 0.6401 | 0.3921 | 0.1863 | 18791 | completed |
| M4_l1_sparsity_only | 50 | 0.2440 | 0.5172 | 0.4054 | 0.1926 | 19502 | completed |
| M5_adaptive_lr_only | 50 | 0.4611 | 0.7419 | 0.3979 | 0.1890 | 19208 | completed |
| M6_eag_only | 50 | 0.2606 | 0.5416 | 0.3992 | 0.1896 | 19319 | completed |
| M7_full_e2am | 50 | 0.5805 | 0.8162 | 0.2005 | 0.0952 | 9683 | completed |
C-matrix (cumulative ablation)
| Variant | Epochs | Top-1 | Top-5 | Energy (kWh) | CO₂ (kg) | Time (s) | Status |
|---|---|---|---|---|---|---|---|
| C0_baseline | 50 | 0.2475 | 0.5199 | 0.4119 | 0.1957 | 19800 | completed |
| C1_cache | 50 | 0.2502 | 0.5186 | 0.4014 | 0.1907 | 19312 | completed |
| C2_cache_amp | 50 | 0.3465 | 0.6300 | 0.1980 | 0.0940 | 9633 | completed |
| C3_cache_amp_gradaccum | 50 | 0.4436 | 0.7242 | 0.2012 | 0.0956 | 9766 | completed |
| C4_cache_amp_gradaccum_adaptivelr | 50 | 0.5738 | 0.8162 | 0.2026 | 0.0963 | 9831 | completed |
| C5_cache_amp_gradaccum_adaptivelr_l1 | 50 | 0.5708 | 0.8125 | 0.2053 | 0.0975 | 9943 | completed |
| C6_full_e2am | 50 | 0.5708 | 0.8125 | 0.2053 | 0.0975 | 9939 | completed |
Deployment results
No deployment results in this repo yet.
Methodology
Model: EfficientNetV2-S (~20.2M params).
Training protocol: from scratch, SGD with momentum 0.9, weight decay 5e-4, initial LR 0.1, 50 epochs, 1 warmup epoch. All variants share the same protocol so ablation comparison stays apples-to-apples across the matrix.
Input: native dataset resolution upsampled to 128x128 in-model via nn.Upsample (FX-traceable to keep D3/D4 INT8 quantization possible).
Optimization toggles (the 5 individual methods and their cumulative combinations):
| Method | Mechanism |
|---|---|
| Tensor cache | Training images held in RAM as a normalized float tensor |
| AMP | torch.cuda.amp.autocast + GradScaler |
| Grad accum (x2) | Accumulate gradients across 2 mini-batches |
| L1 sparsity | Lambda * sum( |
| Cosine LR | lr(t) = lr_max * 0.5 * (1 + cos(pi*t/T)) |
| EAG early-stop | Energy-Aware Gain: stop when accuracy gain per joule plateaus |
Energy measurement: GPU power sampled at 1 Hz via nvidia-smi --query-gpu=power.draw. Energy = trapezoidal integration over power-vs-time. CO₂ = energy_kWh * 0.475 (global average grid intensity).
Hardware: Single NVIDIA T4 (14.5 GB) on Kaggle.
Repository structure
runs/
cifar10/
cifar100/
tiny_imagenet/
individual_methods/M0..M7/ (history.csv, metrics_summary.json,
best_model.pt, last_model.pt, config.yaml)
cumulative_ablation/C0..C6/ (same)
paper_tables/ (6 unified CSV tables)
comparison_plots/<dataset>/ (per-dataset plots)
comparison_plots/cross_dataset/ (cross-dataset plots)
README.md (this file)
Reproducibility
Each variant directory has a config.yaml with the exact configuration used. To reproduce:
huggingface-cli download Shanmuk4622/E2AM_EfficientNetV2_S --repo-type dataset- Load the
e2am.pylibrary and call the appropriate config factory - Run
e2am.train_one_run(cfg)
Limitations
- Energy measurement is GPU-only (via nvidia-smi); CPU/memory power not included
- Pruning is mask-based; no wall-clock speedup without sparsity-aware runtime
- INT8 (D3/D4) is CPU FX static quantization (fbgemm); may fail on transformer blocks. Failures logged in metrics.json rather than crashing.
- Single-T4 reproduction; multi-GPU not validated
- SGD@0.1 is suboptimal for some architectures; the paper compares variant-to-variant deltas which remain meaningful regardless
Citation
@misc{e2am_ablation_effnetv2s,
title = {E2AM: Energy-Aware Adaptive Model Training Ablation Study (EfficientNetV2-S)},
author = {Shanmuk},
year = {2026},
howpublished = {\url{https://huggingface.co/datasets/Shanmuk4622/E2AM_EfficientNetV2_S}},
}
This README was auto-generated on 2026-06-06 12:40 UTC. Source repo: Shanmuk4622/E2AM_EfficientNetV2_S
- Downloads last month
- 292













