assignment-1 run mlp-seed2-20261019-225325

assignment-1/runs/mlp-seed2-20261019-225325/checkpoint.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c7e2ef71f89df2f0146e4274adf1f8c29d16f928c4a975911029e11dfb108191
+size 2146361

assignment-1/runs/mlp-seed2-20261019-225325/checksums.json

@@ -0,0 +1,13 @@
+{
+  "checkpoint.pt": "c7e2ef71f89df2f0146e4274adf1f8c29d16f928c4a975911029e11dfb108191",
+  "config.yaml": "ab4f97c4827d479939d0807d2092da4a3944a30091c1c6fa1c51494072bae9d8",
+  "environment.json": "0e6cb2d99270a3e06a7c1ccc1840746db19f4b623550957e3374d32e83f7ed0a",
+  "history.json": "19756a9c7d524b948bec7c817eb577612d8384b0e5f5c465f6ca51d241aa882d",
+  "summary.json": "409c5f8c0321f18bd54bfd25667cb804c736a56390dcc4634de6dace10ebaeb8",
+  "evaluation/val/correct.png": "3caf25953f965ff9e78ef1d5d82d73487da35b34a3a651d42c806c617be45eaa",
+  "evaluation/val/incorrect.png": "211a0a4ddd179982ba25aa1e3595275d014ea31765c973159e411f36e32b0e6c",
+  "evaluation/val/report.json": "b945e4acb296bd72502d8e512ff7a1f4f0b3f2fcce7869feebf0e1554708a3c2",
+  "evaluation/test/correct.png": "f91b9e601b2e5165b54696933da9c25ef5e1b59b8b1f49f610876c886dc37697",
+  "evaluation/test/incorrect.png": "9b4ddf2fad0d8068cd1069af2687ed54b1dcf970f44ddd020f1fa8b89fd00f7b",
+  "evaluation/test/report.json": "1cd7f53484f01fa969f84504024719544ff9e26b0ed5d463d11706e74ce92acc"
+}

assignment-1/runs/mlp-seed2-20261019-225325/config.yaml

@@ -0,0 +1,55 @@
+seed: 2
+dataset:
+  name: fashion-mnist
+  source: torchvision.datasets.FashionMNIST
+  root: data
+  download: true
+  val_size: 6000
+  stratified: true
+  split_file: splits/fashion-mnist-seed0.json
+  normalization:
+    mean: 0.28579843044281006
+    std: 0.3528536558151245
+hub:
+  repo_id: null
+  assignment: assignment-1
+  private: false
+input:
+  representation: image
+  patch_size: 4
+loader:
+  batch_size: 128
+  num_workers: 4
+  augment: true
+train:
+  epochs: 30
+  optimizer: adamw
+  lr: 0.0003
+  weight_decay: 0.05
+  scheduler: cosine
+  label_smoothing: 0.0
+  early_stopping_patience: 5
+  checkpoint_metric: val_macro_f1
+  checkpoint_mode: max
+evaluation:
+  metrics:
+  - accuracy
+  - macro_f1
+  confusion_matrix: true
+  example_predictions: 8
+  output_dir: results/evaluation
+  inference:
+    batch_sizes:
+    - 1
+    - 128
+    warmup_batches: 5
+    timed_batches: 50
+model:
+  args:
+    hidden_sizes:
+    - 512
+    - 256
+    activation: relu
+    dropout: 0.2
+  name: mlp
+output_dir: results/runs

assignment-1/runs/mlp-seed2-20261019-225325/environment.json

@@ -0,0 +1,10 @@
+{
+  "commit": "42f775dbfcb780f7517ddc8ea8f33863b848da6d",
+  "python": "3.12.0",
+  "torch": "2.14.0",
+  "numpy": "2.5.3",
+  "torchvision": "0.29.0",
+  "scikit_learn": "1.9.1",
+  "hardware": "mps: Apple M4 Pro, 14 cores, 48 GB RAM",
+  "torch_threads": 10
+}

assignment-1/runs/mlp-seed2-20261019-225325/evaluation/test/report.json

@@ -0,0 +1,172 @@
+{
+  "run_id": "mlp-seed2-20261019-225325",
+  "model": "mlp",
+  "representation": "image",
+  "split": "test",
+  "accuracy": 0.8791999816894531,
+  "macro_f1": 0.8790710635253219,
+  "per_class_precision": {
+    "T-shirt/top": 0.827037773359841,
+    "Trouser": 0.9888324873096447,
+    "Pullover": 0.8161993769470405,
+    "Dress": 0.8862512363996043,
+    "Coat": 0.7850287907869482,
+    "Sandal": 0.9545454545454546,
+    "Shirt": 0.6915219611848825,
+    "Sneaker": 0.9186834462729913,
+    "Bag": 0.9710867397806581,
+    "Ankle boot": 0.9524291497975709
+  },
+  "per_class_recall": {
+    "T-shirt/top": 0.832,
+    "Trouser": 0.974,
+    "Pullover": 0.786,
+    "Dress": 0.896,
+    "Coat": 0.818,
+    "Sandal": 0.945,
+    "Shirt": 0.677,
+    "Sneaker": 0.949,
+    "Bag": 0.974,
+    "Ankle boot": 0.941
+  },
+  "per_class_f1": {
+    "T-shirt/top": 0.8295114656031904,
+    "Trouser": 0.981360201511335,
+    "Pullover": 0.8008150789607743,
+    "Dress": 0.8910989557434112,
+    "Coat": 0.801175318315377,
+    "Sandal": 0.949748743718593,
+    "Shirt": 0.6841839312784235,
+    "Sneaker": 0.9335956714215445,
+    "Bag": 0.9725411882176734,
+    "Ankle boot": 0.9466800804828974
+  },
+  "confusion_matrix": [
+    [
+      832,
+      2,
+      11,
+      23,
+      4,
+      3,
+      119,
+      0,
+      6,
+      0
+    ],
+    [
+      3,
+      974,
+      0,
+      16,
+      3,
+      0,
+      2,
+      0,
+      2,
+      0
+    ],
+    [
+      11,
+      1,
+      786,
+      10,
+      104,
+      0,
+      87,
+      0,
+      1,
+      0
+    ],
+    [
+      19,
+      8,
+      7,
+      896,
+      36,
+      1,
+      29,
+      0,
+      4,
+      0
+    ],
+    [
+      1,
+      0,
+      85,
+      33,
+      818,
+      0,
+      61,
+      0,
+      2,
+      0
+    ],
+    [
+      0,
+      0,
+      0,
+      0,
+      0,
+      945,
+      0,
+      37,
+      1,
+      17
+    ],
+    [
+      137,
+      0,
+      72,
+      28,
+      73,
+      0,
+      677,
+      0,
+      13,
+      0
+    ],
+    [
+      0,
+      0,
+      0,
+      0,
+      0,
+      21,
+      0,
+      949,
+      0,
+      30
+    ],
+    [
+      2,
+      0,
+      2,
+      5,
+      4,
+      5,
+      4,
+      4,
+      974,
+      0
+    ],
+    [
+      1,
+      0,
+      0,
+      0,
+      0,
+      15,
+      0,
+      43,
+      0,
+      941
+    ]
+  ],
+  "parameters": 535818,
+  "training_seconds": 351.31143403053284,
+  "inference_ms_per_image": {
+    "1": 0.12834416003897786,
+    "128": 0.002231289061001007
+  }
+}

assignment-1/runs/mlp-seed2-20261019-225325/evaluation/val/report.json

@@ -0,0 +1,172 @@
+{
+  "run_id": "mlp-seed2-20261019-225325",
+  "model": "mlp",
+  "representation": "image",
+  "split": "val",
+  "accuracy": 0.8853333592414856,
+  "macro_f1": 0.884935905627564,
+  "per_class_precision": {
+    "T-shirt/top": 0.8133748055987559,
+    "Trouser": 0.9898477157360406,
+    "Pullover": 0.82842287694974,
+    "Dress": 0.8859934853420195,
+    "Coat": 0.7691107644305772,
+    "Sandal": 0.9633943427620633,
+    "Shirt": 0.728440366972477,
+    "Sneaker": 0.9315960912052117,
+    "Bag": 0.9811643835616438,
+    "Ankle boot": 0.9627118644067797
+  },
+  "per_class_recall": {
+    "T-shirt/top": 0.8716666666666667,
+    "Trouser": 0.975,
+    "Pullover": 0.7966666666666666,
+    "Dress": 0.9066666666666666,
+    "Coat": 0.8216666666666667,
+    "Sandal": 0.965,
+    "Shirt": 0.6616666666666666,
+    "Sneaker": 0.9533333333333334,
+    "Bag": 0.955,
+    "Ankle boot": 0.9466666666666667
+  },
+  "per_class_f1": {
+    "T-shirt/top": 0.8415124698310539,
+    "Trouser": 0.982367758186398,
+    "Pullover": 0.8122344944774851,
+    "Dress": 0.8962108731466227,
+    "Coat": 0.7945205479452054,
+    "Sandal": 0.9641965029142381,
+    "Shirt": 0.6934497816593886,
+    "Sneaker": 0.942339373970346,
+    "Bag": 0.9679054054054054,
+    "Ankle boot": 0.9546218487394958
+  },
+  "confusion_matrix": [
+    [
+      523,
+      0,
+      4,
+      16,
+      2,
+      1,
+      51,
+      0,
+      3,
+      0
+    ],
+    [
+      2,
+      585,
+      0,
+      10,
+      0,
+      0,
+      2,
+      0,
+      1,
+      0
+    ],
+    [
+      5,
+      0,
+      478,
+      6,
+      66,
+      2,
+      42,
+      0,
+      1,
+      0
+    ],
+    [
+      13,
+      6,
+      2,
+      544,
+      26,
+      0,
+      8,
+      0,
+      1,
+      0
+    ],
+    [
+      2,
+      0,
+      48,
+      19,
+      493,
+      0,
+      38,
+      0,
+      0,
+      0
+    ],
+    [
+      0,
+      0,
+      0,
+      1,
+      0,
+      579,
+      0,
+      15,
+      1,
+      4
+    ],
+    [
+      96,
+      0,
+      39,
+      13,
+      51,
+      0,
+      397,
+      0,
+      4,
+      0
+    ],
+    [
+      0,
+      0,
+      0,
+      0,
+      0,
+      11,
+      0,
+      572,
+      0,
+      17
+    ],
+    [
+      2,
+      0,
+      6,
+      5,
+      3,
+      3,
+      7,
+      0,
+      573,
+      1
+    ],
+    [
+      0,
+      0,
+      0,
+      0,
+      0,
+      5,
+      0,
+      27,
+      0,
+      568
+    ]
+  ],
+  "parameters": 535818,
+  "training_seconds": 351.31143403053284,
+  "inference_ms_per_image": {
+    "1": 0.1276024998514913,
+    "128": 0.002255957031138678
+  }
+}

assignment-1/runs/mlp-seed2-20261019-225325/history.json

@@ -0,0 +1,332 @@
+[
+  {
+    "epoch": 1,
+    "train_loss": 0.7942518743585657,
+    "train_accuracy": 0.7072407407407407,
+    "train_macro_f1": 0.7020322786070485,
+    "val_loss": 0.5581596735318501,
+    "val_accuracy": 0.7891666666666667,
+    "val_macro_f1": 0.787672526168737,
+    "seconds": 11.61663007736206,
+    "learning_rate": 0.0003
+  },
+  {
+    "epoch": 2,
+    "train_loss": 0.5894865083694458,
+    "train_accuracy": 0.7787037037037037,
+    "train_macro_f1": 0.7752724263152169,
+    "val_loss": 0.4870946266651154,
+    "val_accuracy": 0.8133333333333334,
+    "val_macro_f1": 0.8088228175506449,
+    "seconds": 11.349257946014404,
+    "learning_rate": 0.00029917828430524096
+  },
+  {
+    "epoch": 3,
+    "train_loss": 0.5390800987084706,
+    "train_accuracy": 0.798537037037037,
+    "train_macro_f1": 0.7957722642328846,
+    "val_loss": 0.4505007440249125,
+    "val_accuracy": 0.8268333333333333,
+    "val_macro_f1": 0.8203230212692126,
+    "seconds": 11.430152893066406,
+    "learning_rate": 0.0002967221401100708
+  },
+  {
+    "epoch": 4,
+    "train_loss": 0.5009111086085991,
+    "train_accuracy": 0.814,
+    "train_macro_f1": 0.8117381526081608,
+    "val_loss": 0.42257103673617047,
+    "val_accuracy": 0.8401666666666666,
+    "val_macro_f1": 0.8352579994612744,
+    "seconds": 12.203058958053589,
+    "learning_rate": 0.00029265847744427303
+  },
+  {
+    "epoch": 5,
+    "train_loss": 0.4828034057793794,
+    "train_accuracy": 0.8220185185185185,
+    "train_macro_f1": 0.8198993563362104,
+    "val_loss": 0.4071782606442769,
+    "val_accuracy": 0.845,
+    "val_macro_f1": 0.8451423877978638,
+    "seconds": 11.797888040542603,
+    "learning_rate": 0.0002870318186463901
+  },
+  {
+    "epoch": 6,
+    "train_loss": 0.4621826006394845,
+    "train_accuracy": 0.8293888888888888,
+    "train_macro_f1": 0.8274631341397672,
+    "val_loss": 0.4016587824821472,
+    "val_accuracy": 0.8478333333333333,
+    "val_macro_f1": 0.8428131937157444,
+    "seconds": 11.64467978477478,
+    "learning_rate": 0.0002799038105676658
+  },
+  {
+    "epoch": 7,
+    "train_loss": 0.44967773968202096,
+    "train_accuracy": 0.8336666666666667,
+    "train_macro_f1": 0.8320573129610567,
+    "val_loss": 0.3730386614004771,
+    "val_accuracy": 0.8635,
+    "val_macro_f1": 0.8638995513956631,
+    "seconds": 11.753312110900879,
+    "learning_rate": 0.0002713525491562421
+  },
+  {
+    "epoch": 8,
+    "train_loss": 0.4408185476903562,
+    "train_accuracy": 0.8359444444444445,
+    "train_macro_f1": 0.8344712899638094,
+    "val_loss": 0.37757578722635904,
+    "val_accuracy": 0.8601666666666666,
+    "val_macro_f1": 0.8596764714424026,
+    "seconds": 11.567687034606934,
+    "learning_rate": 0.00026147172382160914
+  },
+  {
+    "epoch": 9,
+    "train_loss": 0.43141501420515554,
+    "train_accuracy": 0.8413333333333334,
+    "train_macro_f1": 0.8397528408909795,
+    "val_loss": 0.3697422614097595,
+    "val_accuracy": 0.8623333333333333,
+    "val_macro_f1": 0.8613650303668914,
+    "seconds": 11.354737281799316,
+    "learning_rate": 0.00025036959095382875
+  },
+  {
+    "epoch": 10,
+    "train_loss": 0.42139923163696574,
+    "train_accuracy": 0.8456111111111111,
+    "train_macro_f1": 0.8442745334927352,
+    "val_loss": 0.35584757685661317,
+    "val_accuracy": 0.8666666666666667,
+    "val_macro_f1": 0.8654748623529901,
+    "seconds": 11.666413068771362,
+    "learning_rate": 0.000238167787843871
+  },
+  {
+    "epoch": 11,
+    "train_loss": 0.41548373824578744,
+    "train_accuracy": 0.8472037037037037,
+    "train_macro_f1": 0.8459697622567072,
+    "val_loss": 0.35199051260948183,
+    "val_accuracy": 0.8675,
+    "val_macro_f1": 0.865308062977715,
+    "seconds": 11.31458306312561,
+    "learning_rate": 0.00022500000000000002
+  },
+  {
+    "epoch": 12,
+    "train_loss": 0.4080266208736985,
+    "train_accuracy": 0.8495555555555555,
+    "train_macro_f1": 0.8483541040236116,
+    "val_loss": 0.34880601946512857,
+    "val_accuracy": 0.8711666666666666,
+    "val_macro_f1": 0.8702410584765106,
+    "seconds": 11.311727046966553,
+    "learning_rate": 0.0002110104964613701
+  },
+  {
+    "epoch": 13,
+    "train_loss": 0.4021190181838142,
+    "train_accuracy": 0.8507222222222223,
+    "train_macro_f1": 0.8495932238866624,
+    "val_loss": 0.3444989873568217,
+    "val_accuracy": 0.8736666666666667,
+    "val_macro_f1": 0.8729001512428548,
+    "seconds": 11.327713251113892,
+    "learning_rate": 0.00019635254915624216
+  },
+  {
+    "epoch": 14,
+    "train_loss": 0.39677745184191954,
+    "train_accuracy": 0.8522037037037037,
+    "train_macro_f1": 0.8510292828734922,
+    "val_loss": 0.33638866074879964,
+    "val_accuracy": 0.8785,
+    "val_macro_f1": 0.8775037462857942,
+    "seconds": 11.315073013305664,
+    "learning_rate": 0.00018118675362266393
+  },
+  {
+    "epoch": 15,
+    "train_loss": 0.39121161063512166,
+    "train_accuracy": 0.8549814814814815,
+    "train_macro_f1": 0.8540301856041476,
+    "val_loss": 0.337692596356074,
+    "val_accuracy": 0.8731666666666666,
+    "val_macro_f1": 0.8730187691625465,
+    "seconds": 11.337677001953125,
+    "learning_rate": 0.0001656792694901481
+  },
+  {
+    "epoch": 16,
+    "train_loss": 0.38397134699644864,
+    "train_accuracy": 0.8571111111111112,
+    "train_macro_f1": 0.8560639030803319,
+    "val_loss": 0.32963196833928426,
+    "val_accuracy": 0.878,
+    "val_macro_f1": 0.8776646266098522,
+    "seconds": 11.619801044464111,
+    "learning_rate": 0.0001500000000000001
+  },
+  {
+    "epoch": 17,
+    "train_loss": 0.37877878984698543,
+    "train_accuracy": 0.8592962962962963,
+    "train_macro_f1": 0.8582884493859971,
+    "val_loss": 0.3306595858732859,
+    "val_accuracy": 0.877,
+    "val_macro_f1": 0.8764535393933022,
+    "seconds": 11.750208139419556,
+    "learning_rate": 0.00013432073050985205
+  },
+  {
+    "epoch": 18,
+    "train_loss": 0.3764293287595113,
+    "train_accuracy": 0.8603148148148149,
+    "train_macro_f1": 0.8593989009915599,
+    "val_loss": 0.3219764742851257,
+    "val_accuracy": 0.8808333333333334,
+    "val_macro_f1": 0.8803560735717249,
+    "seconds": 12.135315895080566,
+    "learning_rate": 0.00011881324637733612
+  },
+  {
+    "epoch": 19,
+    "train_loss": 0.37145271649184053,
+    "train_accuracy": 0.8626481481481482,
+    "train_macro_f1": 0.8617571941890059,
+    "val_loss": 0.31925848849614463,
+    "val_accuracy": 0.8818333333333334,
+    "val_macro_f1": 0.8814481437055802,
+    "seconds": 11.81864619255066,
+    "learning_rate": 0.00010364745084375793
+  },
+  {
+    "epoch": 20,
+    "train_loss": 0.37021037813469215,
+    "train_accuracy": 0.8636111111111111,
+    "train_macro_f1": 0.8628027185159798,
+    "val_loss": 0.31749693767229714,
+    "val_accuracy": 0.8831666666666667,
+    "val_macro_f1": 0.8825955131992901,
+    "seconds": 11.706556797027588,
+    "learning_rate": 8.898950353863002e-05
+  },
+  {
+    "epoch": 21,
+    "train_loss": 0.3632688612673018,
+    "train_accuracy": 0.8656851851851852,
+    "train_macro_f1": 0.864725561870725,
+    "val_loss": 0.3196996295452118,
+    "val_accuracy": 0.881,
+    "val_macro_f1": 0.8809316117171276,
+    "seconds": 11.973679780960083,
+    "learning_rate": 7.500000000000005e-05
+  },
+  {
+    "epoch": 22,
+    "train_loss": 0.35909953461752997,
+    "train_accuracy": 0.867,
+    "train_macro_f1": 0.8661027988958301,
+    "val_loss": 0.3139388559659322,
+    "val_accuracy": 0.8818333333333334,
+    "val_macro_f1": 0.881495314745553,
+    "seconds": 11.91454529762268,
+    "learning_rate": 6.183221215612906e-05
+  },
+  {
+    "epoch": 23,
+    "train_loss": 0.35692365203080356,
+    "train_accuracy": 0.8688148148148148,
+    "train_macro_f1": 0.8680480297020431,
+    "val_loss": 0.3139564705689748,
+    "val_accuracy": 0.8826666666666667,
+    "val_macro_f1": 0.8821340344440515,
+    "seconds": 11.388498783111572,
+    "learning_rate": 4.963040904617133e-05
+  },
+  {
+    "epoch": 24,
+    "train_loss": 0.35673015432887606,
+    "train_accuracy": 0.8687777777777778,
+    "train_macro_f1": 0.8679927123502245,
+    "val_loss": 0.3131129310131073,
+    "val_accuracy": 0.8825,
+    "val_macro_f1": 0.8822762595863545,
+    "seconds": 12.509377002716064,
+    "learning_rate": 3.852827617839091e-05
+  },
+  {
+    "epoch": 25,
+    "train_loss": 0.3524528222702168,
+    "train_accuracy": 0.8690555555555556,
+    "train_macro_f1": 0.8683122883878627,
+    "val_loss": 0.3095294642448425,
+    "val_accuracy": 0.8843333333333333,
+    "val_macro_f1": 0.8838576452240593,
+    "seconds": 12.264047145843506,
+    "learning_rate": 2.8647450843757908e-05
+  },
+  {
+    "epoch": 26,
+    "train_loss": 0.34940404111367684,
+    "train_accuracy": 0.8712037037037037,
+    "train_macro_f1": 0.8704058426504272,
+    "val_loss": 0.3098871602217356,
+    "val_accuracy": 0.8835,
+    "val_macro_f1": 0.8834240118909081,
+    "seconds": 11.51358699798584,
+    "learning_rate": 2.00961894323342e-05
+  },
+  {
+    "epoch": 27,
+    "train_loss": 0.3505620844629076,
+    "train_accuracy": 0.8723148148148148,
+    "train_macro_f1": 0.8716385410053743,
+    "val_loss": 0.30845420773824056,
+    "val_accuracy": 0.8853333333333333,
+    "val_macro_f1": 0.884935905627564,
+    "seconds": 12.350613117218018,
+    "learning_rate": 1.2968181353609857e-05
+  },
+  {
+    "epoch": 28,
+    "train_loss": 0.3493796648890884,
+    "train_accuracy": 0.8707037037037038,
+    "train_macro_f1": 0.8699147153976012,
+    "val_loss": 0.3078193929195404,
+    "val_accuracy": 0.884,
+    "val_macro_f1": 0.8835829858837381,
+    "seconds": 11.819926023483276,
+    "learning_rate": 7.341522555726972e-06
+  },
+  {
+    "epoch": 29,
+    "train_loss": 0.3501284035400108,
+    "train_accuracy": 0.8708703703703704,
+    "train_macro_f1": 0.8700674506880677,
+    "val_loss": 0.30778288197517395,
+    "val_accuracy": 0.8841666666666667,
+    "val_macro_f1": 0.883609368022317,
+    "seconds": 11.718166828155518,
+    "learning_rate": 3.2778598899291474e-06
+  },
+  {
+    "epoch": 30,
+    "train_loss": 0.3484793444121325,
+    "train_accuracy": 0.8706111111111111,
+    "train_macro_f1": 0.8697843869432654,
+    "val_loss": 0.30774388440450035,
+    "val_accuracy": 0.8848333333333334,
+    "val_macro_f1": 0.8843325417619712,
+    "seconds": 11.77487325668335,
+    "learning_rate": 8.2171569475899e-07
+  }
+]

assignment-1/runs/mlp-seed2-20261019-225325/summary.json

@@ -0,0 +1,14 @@
+{
+  "run_id": "mlp-seed2-20261019-225325",
+  "model": "mlp",
+  "representation": "image",
+  "seed": 2,
+  "parameters": 535818,
+  "epochs_trained": 30,
+  "best_epoch": 27,
+  "checkpoint_metric": "val_macro_f1",
+  "checkpoint_value": 0.884935905627564,
+  "training_seconds": 351.31143403053284,
+  "commit": "42f775dbfcb780f7517ddc8ea8f33863b848da6d",
+  "hardware": "mps: Apple M4 Pro, 14 cores, 48 GB RAM"
+}