e2am: batch upload 13 files (322550 KB)
Browse files- runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/best_model.pt +3 -0
- runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/checkpoints/checkpoint_last.pt +3 -0
- runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/config.yaml +33 -0
- runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/figures/accuracy_curve.png +3 -0
- runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/figures/accuracy_vs_energy.png +3 -0
- runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/figures/cumulative_energy.png +3 -0
- runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/figures/epoch_energy.png +3 -0
- runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/figures/gpu_power_trace.png +3 -0
- runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/figures/loss_curve.png +3 -0
- runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/final_model.pt +3 -0
- runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/gpu_power_trace.csv +0 -0
- runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/history.csv +9 -0
- runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/metrics_summary.json +27 -0
runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/best_model.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:195540427f65ef2c4ef7c60d9f9ffb04dbf8282f2f1c85659e7391c0e70ba045
|
| 3 |
+
size 82657563
|
runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/checkpoints/checkpoint_last.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:451cade2d65c864d45a28588d892db6e6c20a880b5e3f12ca3732317d243b6d0
|
| 3 |
+
size 164532813
|
runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/config.yaml
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
project_name: E2AM-ResNet50
|
| 2 |
+
model_name: effnetv2_s
|
| 3 |
+
dataset_name: tiny_imagenet
|
| 4 |
+
output_root: /kaggle/working/E2AM-EffNetV2-TIN
|
| 5 |
+
data_root: /kaggle/working/tiny-imagenet-200
|
| 6 |
+
seed: 42
|
| 7 |
+
num_epochs: 50
|
| 8 |
+
batch_size: 128
|
| 9 |
+
image_size: 32
|
| 10 |
+
optimizer: sgd
|
| 11 |
+
learning_rate: 0.1
|
| 12 |
+
momentum: 0.9
|
| 13 |
+
weight_decay: 0.0005
|
| 14 |
+
warmup_epochs: 2
|
| 15 |
+
amp_enabled: true
|
| 16 |
+
cache_dataset: true
|
| 17 |
+
gradient_accumulation_steps: 2
|
| 18 |
+
l1_lambda: 1.0e-08
|
| 19 |
+
scheduler: cosine
|
| 20 |
+
eag_enabled: false
|
| 21 |
+
eag_threshold: 1.0e-07
|
| 22 |
+
eag_patience: 5
|
| 23 |
+
energy_sample_interval_sec: 1.0
|
| 24 |
+
carbon_intensity_kg_per_kwh: 0.475
|
| 25 |
+
milestone_push_every_epochs: 10
|
| 26 |
+
heavy_push_every_epochs: 10
|
| 27 |
+
force_rerun: false
|
| 28 |
+
cleanup_local_after_complete: true
|
| 29 |
+
cleanup_keep_summary: false
|
| 30 |
+
model_input_size: 128
|
| 31 |
+
method_group: cumulative_ablation
|
| 32 |
+
variant_name: C5_cache_amp_gradaccum_adaptivelr_l1
|
| 33 |
+
method_id: C5
|
runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/figures/accuracy_curve.png
ADDED
|
Git LFS Details
|
runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/figures/accuracy_vs_energy.png
ADDED
|
Git LFS Details
|
runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/figures/cumulative_energy.png
ADDED
|
Git LFS Details
|
runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/figures/epoch_energy.png
ADDED
|
Git LFS Details
|
runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/figures/gpu_power_trace.png
ADDED
|
Git LFS Details
|
runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/figures/loss_curve.png
ADDED
|
Git LFS Details
|
runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/final_model.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2fe1d9e9cdacc00ffc4bf0e7c51f1399167ebda2282a58c3a1b4679fc5e162a6
|
| 3 |
+
size 82658351
|
runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/gpu_power_trace.csv
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/history.csv
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
epoch,train_loss,val_loss,train_accuracy,val_accuracy,val_accuracy_top5,f1_score,precision,recall,learning_rate,batch_size,effective_batch_size,amp_enabled,l1_lambda,gradient_accumulation_steps,epoch_time_sec,cumulative_time_sec,epoch_energy_j,cumulative_energy_j,epoch_co2_kg,cumulative_co2_kg,peak_vram_mb,eag_score,early_stop_counter,timestamp_utc
|
| 2 |
+
0,5.191445872802734,4.81088173828125,0.01643,0.0406,0.1495,0.020130561901125034,0.021820298776702482,0.040600000000000004,0.05,128,256,True,1e-08,2,200.19654321670532,200.19654321670532,14853.367777053196,14853.367777053196,0.0019598193594722963,0.0019598193594722963,3518.37255859375,0.0,0,2026-05-20T19:04:29Z
|
| 3 |
+
1,4.580623174438476,4.354086938476563,0.06287,0.0858,0.2555,0.06596892748406642,0.08551321255386977,0.0858,0.1,128,256,True,1e-08,2,211.72483468055725,411.9213778972626,15517.557645411134,30370.92542246433,0.0020474555226584136,0.00400727488213071,3518.37255859375,2.9128295207826464e-06,0,2026-05-20T19:08:08Z
|
| 4 |
+
2,4.041269759216308,3.784387548828125,0.12795,0.1618,0.3939,0.13702260407163278,0.1693312446958154,0.16180000000000003,0.09990133642141358,128,256,True,1e-08,2,200.33207988739014,612.2534577846527,14900.657855649311,45271.58327811364,0.001966059022620395,0.005973333904751106,3518.37255859375,5.100445949182437e-06,0,2026-05-20T19:11:38Z
|
| 5 |
+
3,3.674536591796875,3.675132489013672,0.18085,0.1858,0.4284,0.16648304269910072,0.20417876417287664,0.18580000000000002,0.0996057350657239,128,256,True,1e-08,2,200.24147510528564,812.4949328899384,14906.268912198359,60177.852190312,0.0019667993703595056,0.007940133275110612,3518.37255859375,1.6100608503285415e-06,0,2026-05-20T19:15:08Z
|
| 6 |
+
4,3.459441448059082,3.404514495849609,0.21754,0.2295,0.4922,0.20444400417050468,0.2401205939061814,0.22950000000000004,0.09911436253643444,128,256,True,1e-08,2,200.57471632957458,1013.0696492195129,14862.831481875819,75040.68367218782,0.0019610680427475037,0.009901201317858115,3518.37255859375,2.940220378148612e-06,0,2026-05-20T19:18:39Z
|
| 7 |
+
5,3.3444943185424805,3.3207004089355467,0.23757,0.2409,0.5026,0.21818820242641945,0.2576587060896048,0.24090000000000003,0.09842915805643154,128,256,True,1e-08,2,200.44020915031433,1213.5098583698273,14858.376581329218,89899.06025351703,0.001960480243369827,0.011861681561227942,3518.37255859375,7.672439810365985e-07,0,2026-05-20T19:22:09Z
|
| 8 |
+
6,3.2234693389892577,3.232755126953125,0.26002,0.2642,0.5403,0.24100430398786402,0.27649739926493555,0.2642,0.09755282581475767,128,256,True,1e-08,2,200.49744081497192,1414.0072991847992,15028.523609315587,104927.58386283262,0.001982930198451362,0.013844611759679304,3518.37255859375,1.5503851612913752e-06,0,2026-05-20T19:25:39Z
|
| 9 |
+
7,3.132363767700195,3.208836083984375,0.27554,0.2796,0.5627,0.2645714057057943,0.30477509871492603,0.2796,0.09648882429441256,128,256,True,1e-08,2,200.61828875541687,1614.625587940216,14981.997738548072,119909.58160138069,0.001976791368280648,0.015821403127959953,3518.37255859375,1.027900301998875e-06,0,2026-05-20T19:29:09Z
|
runs/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1/metrics_summary.json
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"run_key": "effnetv2_s::tiny_imagenet::cumulative_ablation::C5_cache_amp_gradaccum_adaptivelr_l1",
|
| 3 |
+
"run_dir": "/kaggle/working/E2AM-EffNetV2-TIN/tiny_imagenet/cumulative_ablation/C5_cache_amp_gradaccum_adaptivelr_l1",
|
| 4 |
+
"model_name": "effnetv2_s",
|
| 5 |
+
"dataset_name": "tiny_imagenet",
|
| 6 |
+
"method_group": "cumulative_ablation",
|
| 7 |
+
"variant_name": "C5_cache_amp_gradaccum_adaptivelr_l1",
|
| 8 |
+
"method_id": "C5",
|
| 9 |
+
"num_epochs_planned": 50,
|
| 10 |
+
"num_epochs_run_so_far": 8,
|
| 11 |
+
"best_accuracy": 0.2796,
|
| 12 |
+
"latest_val_accuracy": 0.2796,
|
| 13 |
+
"latest_f1_score": 0.2645714057057943,
|
| 14 |
+
"total_time_sec_so_far": 1614.625587940216,
|
| 15 |
+
"total_energy_j_so_far": 119909.58160138069,
|
| 16 |
+
"total_co2_kg_so_far": 0.015821403127959953,
|
| 17 |
+
"peak_vram_mb": 3518.37255859375,
|
| 18 |
+
"amp_enabled": true,
|
| 19 |
+
"scheduler": "cosine",
|
| 20 |
+
"gradient_accumulation_steps": 2,
|
| 21 |
+
"l1_lambda": 1e-08,
|
| 22 |
+
"eag_enabled": false,
|
| 23 |
+
"batch_size": 128,
|
| 24 |
+
"image_size": 32,
|
| 25 |
+
"status": "running",
|
| 26 |
+
"updated_utc": "2026-05-20T19:29:10Z"
|
| 27 |
+
}
|