From 36c403937df1742849acc0db14607d271ca5dfe7 Mon Sep 17 00:00:00 2001
From: "github-classroom[bot]"
<66690702+github-classroom[bot]@users.noreply.github.com>
Date: Mon, 29 Sep 2025 00:42:31 +0000
Subject: [PATCH 1/8] GitHub Classroom Autograding Workflow
---
.github/workflows/classroom.yml | 69 +++++++++++++++++++++++++++++++++
1 file changed, 69 insertions(+)
create mode 100644 .github/workflows/classroom.yml
diff --git a/.github/workflows/classroom.yml b/.github/workflows/classroom.yml
new file mode 100644
index 00000000..e7879f7d
--- /dev/null
+++ b/.github/workflows/classroom.yml
@@ -0,0 +1,69 @@
+name: Autograding Tests
+'on':
+- workflow_dispatch
+- repository_dispatch
+permissions:
+ checks: write
+ actions: read
+ contents: read
+jobs:
+ run-autograding-tests:
+ runs-on: ubuntu-latest
+ if: github.actor != 'github-classroom[bot]'
+ steps:
+ - name: Checkout code
+ uses: actions/checkout@v4
+ - name: Setup
+ id: setup
+ uses: classroom-resources/autograding-command-grader@v1
+ with:
+ test-name: Setup
+ setup-command: sudo -H pip3 install -qr requirements.txt; sudo -H pip3 install
+ flake8==5.0.4
+ command: flake8 --ignore "N801, E203, E266, E501, W503, F812, E741, N803,
+ N802, N806" minitorch/ tests/ project/; mypy minitorch/*
+ timeout: 10
+ max-score: 10
+ - name: Task 1.1
+ id: task-1-1
+ uses: classroom-resources/autograding-command-grader@v1
+ with:
+ test-name: Task 1.1
+ setup-command: sudo -H pip3 install -qr requirements.txt
+ command: pytest -m task1_1
+ timeout: 10
+ max-score: 10
+ - name: Task 1.2
+ id: task-1-2
+ uses: classroom-resources/autograding-command-grader@v1
+ with:
+ test-name: Task 1.2
+ setup-command: sudo -H pip3 install -qr requirements.txt
+ command: pytest -m task1_2
+ timeout: 10
+ - name: task 1.3
+ id: task-1-3
+ uses: classroom-resources/autograding-command-grader@v1
+ with:
+ test-name: task 1.3
+ setup-command: sudo -H pip3 install -qr requirements.txt
+ command: pytest -m task1_3
+ timeout: 10
+ - name: Task 1.4
+ id: task-1-4
+ uses: classroom-resources/autograding-command-grader@v1
+ with:
+ test-name: Task 1.4
+ setup-command: sudo -H pip3 install -qr requirements.txt
+ command: pytest -m task1_4
+ timeout: 10
+ - name: Autograding Reporter
+ uses: classroom-resources/autograding-grading-reporter@v1
+ env:
+ SETUP_RESULTS: "${{steps.setup.outputs.result}}"
+ TASK-1-1_RESULTS: "${{steps.task-1-1.outputs.result}}"
+ TASK-1-2_RESULTS: "${{steps.task-1-2.outputs.result}}"
+ TASK-1-3_RESULTS: "${{steps.task-1-3.outputs.result}}"
+ TASK-1-4_RESULTS: "${{steps.task-1-4.outputs.result}}"
+ with:
+ runners: setup,task-1-1,task-1-2,task-1-3,task-1-4
From 5d5efbebbadbf8b1f01290f8622a94cde0b4c5b7 Mon Sep 17 00:00:00 2001
From: "github-classroom[bot]"
<66690702+github-classroom[bot]@users.noreply.github.com>
Date: Mon, 29 Sep 2025 00:42:31 +0000
Subject: [PATCH 2/8] GitHub Classroom Feedback
---
.github/.keep | 0
1 file changed, 0 insertions(+), 0 deletions(-)
create mode 100644 .github/.keep
diff --git a/.github/.keep b/.github/.keep
new file mode 100644
index 00000000..e69de29b
From 846b159ea788d85942959570d6ba31d6d2c3f506 Mon Sep 17 00:00:00 2001
From: "github-classroom[bot]"
<66690702+github-classroom[bot]@users.noreply.github.com>
Date: Mon, 29 Sep 2025 00:42:32 +0000
Subject: [PATCH 3/8] Setting up GitHub Classroom Feedback
From 0bd00b0e457a5506c30f8170dcd009633dc7e1eb Mon Sep 17 00:00:00 2001
From: "github-classroom[bot]"
<66690702+github-classroom[bot]@users.noreply.github.com>
Date: Mon, 29 Sep 2025 00:42:34 +0000
Subject: [PATCH 4/8] add online IDE url
---
README.md | 1 +
1 file changed, 1 insertion(+)
diff --git a/README.md b/README.md
index 46933775..ba4242fa 100644
--- a/README.md
+++ b/README.md
@@ -1,3 +1,4 @@
+[](https://classroom.github.com/online_ide?assignment_repo_id=20763289&assignment_repo_type=AssignmentRepo)
# MiniTorch Module 1
From a60f2eed59488a551364aac7672ffc2de5037ce6 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=D0=92=D0=B0=D0=B4=D0=B8=D0=BC=20=D0=9F=D0=B0=D0=BB=D1=8B?=
=?UTF-8?q?=D1=81=D0=B0=D0=B5=D0=B2?=
Date: Tue, 30 Sep 2025 23:48:33 +0300
Subject: [PATCH 5/8] m2
---
README.md | 661 ++++++++++++++++++++++++++++++++++
logs/circle.txt | Bin 0 -> 11788 bytes
logs/diag.txt | Bin 0 -> 7862 bytes
logs/simple.txt | Bin 0 -> 8456 bytes
logs/spiral.txt | Bin 0 -> 14524 bytes
logs/split.txt | Bin 0 -> 7812 bytes
logs/xor.txt | Bin 0 -> 9750 bytes
minitorch/autodiff.py | 46 ++-
minitorch/module.py | 56 +--
minitorch/operators.py | 203 +++++------
minitorch/scalar.py | 39 +-
minitorch/scalar_functions.py | 62 ++--
project/run_scalar.py | 23 +-
13 files changed, 888 insertions(+), 202 deletions(-)
create mode 100644 logs/circle.txt
create mode 100644 logs/diag.txt
create mode 100644 logs/simple.txt
create mode 100644 logs/spiral.txt
create mode 100644 logs/split.txt
create mode 100644 logs/xor.txt
diff --git a/README.md b/README.md
index ba4242fa..933b4849 100644
--- a/README.md
+++ b/README.md
@@ -16,3 +16,664 @@ python sync_previous_module.py previous-module-dir current-module-dir
The files that will be synced are:
minitorch/operators.py minitorch/module.py tests/test_module.py tests/test_operators.py project/run_manual.py
+
+
+# Scalar Training
+
+Ниже результаты обучения сети на разных наборах данных:
+
+## Simple
+**Config:** `PTS=50, HIDDEN=2, RATE=0.5`
+
+**Logs (excerpt):**
+
+Epoch 10 loss 32.99451624777131 correct 29
+Epoch 20 loss 29.67106517619935 correct 44
+Epoch 30 loss 20.505422569195055 correct 49
+Epoch 40 loss 12.040371792254232 correct 49
+Epoch 50 loss 8.05779782525025 correct 50
+Epoch 60 loss 5.776877152171056 correct 50
+Epoch 70 loss 4.32599156373445 correct 50
+Epoch 80 loss 3.391007016923641 correct 50
+Epoch 90 loss 2.7390332746561117 correct 50
+Epoch 100 loss 2.2638205155347118 correct 50
+Epoch 110 loss 1.9058296475262912 correct 50
+Epoch 130 loss 1.4126621296878168 correct 50
+Epoch 140 loss 1.2384903100021274 correct 50
+Epoch 140 loss 1.2384903100021274 correct 50
+Epoch 150 loss 1.0965209310276882 correct 50
+Epoch 150 loss 1.0965209310276882 correct 50
+Epoch 160 loss 0.9791433120667529 correct 50
+Epoch 170 loss 0.8810811523651478 correct 50
+Epoch 160 loss 0.9791433120667529 correct 50
+Epoch 170 loss 0.8810811523651478 correct 50
+Epoch 180 loss 0.798041390403876 correct 50
+Epoch 190 loss 0.7270314986517517 correct 50
+Epoch 180 loss 0.798041390403876 correct 50
+Epoch 190 loss 0.7270314986517517 correct 50
+Epoch 200 loss 0.6657803100113434 correct 50
+Epoch 210 loss 0.6127190836144141 correct 50
+Epoch 200 loss 0.6657803100113434 correct 50
+Epoch 210 loss 0.6127190836144141 correct 50
+Epoch 220 loss 0.5666541785345667 correct 50
+Epoch 220 loss 0.5666541785345667 correct 50
+Epoch 230 loss 0.5259841239283145 correct 50
+Epoch 240 loss 0.4899498393389119 correct 50
+Epoch 230 loss 0.5259841239283145 correct 50
+Epoch 240 loss 0.4899498393389119 correct 50
+Epoch 250 loss 0.4580529051584932 correct 50
+Epoch 260 loss 0.42947990336786396 correct 50
+Epoch 250 loss 0.4580529051584932 correct 50
+Epoch 260 loss 0.42947990336786396 correct 50
+Epoch 270 loss 0.40371825574841774 correct 50
+Epoch 280 loss 0.3803927096251919 correct 50
+Epoch 270 loss 0.40371825574841774 correct 50
+Epoch 280 loss 0.3803927096251919 correct 50
+Epoch 290 loss 0.3591923758308195 correct 50
+Epoch 300 loss 0.3398567923406658 correct 50
+Epoch 290 loss 0.3591923758308195 correct 50
+Epoch 300 loss 0.3398567923406658 correct 50
+Epoch 310 loss 0.32216577622298964 correct 50
+Epoch 320 loss 0.305931791725392 correct 50
+Epoch 310 loss 0.32216577622298964 correct 50
+Epoch 320 loss 0.305931791725392 correct 50
+Epoch 330 loss 0.29099407053442744 correct 50
+Epoch 330 loss 0.29099407053442744 correct 50
+Epoch 340 loss 0.2772139984633007 correct 50
+Epoch 350 loss 0.2644714439861456 correct 50
+Epoch 340 loss 0.2772139984633007 correct 50
+Epoch 350 loss 0.2644714439861456 correct 50
+Epoch 360 loss 0.25266180270088984 correct 50
+Epoch 350 loss 0.2644714439861456 correct 50
+Epoch 360 loss 0.25266180270088984 correct 50
+Epoch 360 loss 0.25266180270088984 correct 50
+Epoch 370 loss 0.2416935953168232 correct 50
+Epoch 380 loss 0.23148649944459582 correct 50
+Epoch 390 loss 0.22196972517478103 correct 50
+Epoch 390 loss 0.22196972517478103 correct 50
+Epoch 400 loss 0.2130806656978178 correct 50
+Epoch 400 loss 0.2130806656978178 correct 50
+Epoch 410 loss 0.20476376978405275 correct 50
+Epoch 410 loss 0.20476376978405275 correct 50
+Epoch 420 loss 0.19696959454012272 correct 50
+Epoch 420 loss 0.19696959454012272 correct 50
+Epoch 430 loss 0.18965400562421794 correct 50
+Epoch 430 loss 0.18965400562421794 correct 50
+Epoch 440 loss 0.18277749880928493 correct 50
+Epoch 440 loss 0.18277749880928493 correct 50
+Epoch 450 loss 0.17630462196666688 correct 50
+Epoch 460 loss 0.17020348058519222 correct 50
+Epoch 470 loss 0.16444531311832322 correct 50
+Epoch 450 loss 0.17630462196666688 correct 50
+Epoch 460 loss 0.17020348058519222 correct 50
+Epoch 470 loss 0.16444531311832322 correct 50
+Epoch 460 loss 0.17020348058519222 correct 50
+Epoch 470 loss 0.16444531311832322 correct 50
+Epoch 470 loss 0.16444531311832322 correct 50
+Epoch 480 loss 0.15900412496832433 correct 50
+Epoch 490 loss 0.15385637192186644 correct 50
+Epoch 500 loss 0.14898068545998813 correct 50
+
+
+## Diag
+**Config:** `PTS=50, HIDDEN=3, RATE=0.5, EPOCHS=800`
+
+**Logs:**
+
+Epoch 10 loss 18.41767648121497 correct 44
+Epoch 20 loss 18.241374745482055 correct 44
+Epoch 30 loss 18.077694322397104 correct 44
+Epoch 40 loss 17.787496076171966 correct 44
+Epoch 50 loss 17.22622063447276 correct 44
+Epoch 60 loss 16.218299532974434 correct 44
+Epoch 70 loss 14.60746253340782 correct 44
+Epoch 80 loss 12.23724317791127 correct 44
+Epoch 90 loss 9.544850994180615 correct 46
+Epoch 100 loss 7.562065011671856 correct 46
+Epoch 110 loss 6.112602137050321 correct 48
+Epoch 120 loss 5.0048459933990905 correct 48
+Epoch 130 loss 4.162684014938122 correct 48
+Epoch 140 loss 3.52155744883408 correct 50
+Epoch 150 loss 3.0258024023642136 correct 50
+Epoch 160 loss 2.7067221081928436 correct 50
+Epoch 170 loss 2.982213947123389 correct 50
+Epoch 180 loss 3.571243230756708 correct 49
+Epoch 190 loss 3.085185630314938 correct 50
+Epoch 200 loss 2.6323532616598326 correct 50
+Epoch 210 loss 2.2357765540748717 correct 50
+Epoch 220 loss 1.7922072530790387 correct 50
+Epoch 230 loss 1.4067375619879097 correct 50
+Epoch 240 loss 1.2235048686696837 correct 50
+Epoch 250 loss 1.1016726862974993 correct 50
+Epoch 260 loss 1.0035961711498014 correct 50
+Epoch 270 loss 0.9205055811305474 correct 50
+Epoch 280 loss 0.8454999986534688 correct 50
+Epoch 290 loss 0.7841042944912567 correct 50
+Epoch 300 loss 0.729316067557551 correct 50
+Epoch 310 loss 0.6798828131913032 correct 50
+Epoch 320 loss 0.6351716025665827 correct 50
+Epoch 330 loss 0.5946362524099438 correct 50
+Epoch 340 loss 0.5578028898126859 correct 50
+Epoch 350 loss 0.5242590057316765 correct 50
+Epoch 360 loss 0.4936445178733028 correct 50
+Epoch 370 loss 0.46564435582316355 correct 50
+Epoch 380 loss 0.4399822374613626 correct 50
+Epoch 390 loss 0.41641540204658256 correct 50
+Epoch 400 loss 0.39473012676172964 correct 50
+Epoch 410 loss 0.37473789411177993 correct 50
+Epoch 420 loss 0.3562721055229054 correct 50
+Epoch 430 loss 0.3391852565769399 correct 50
+Epoch 440 loss 0.3233465043336502 correct 50
+Epoch 450 loss 0.3086395688231362 correct 50
+Epoch 460 loss 0.29496092005563224 correct 50
+Epoch 470 loss 0.28221820943708226 correct 50
+Epoch 480 loss 0.2703289107133867 correct 50
+Epoch 490 loss 0.2592191407757518 correct 50
+Epoch 500 loss 0.24882263504409974 correct 50
+Epoch 510 loss 0.23907985585290037 correct 50
+Epoch 520 loss 0.22993721540795956 correct 50
+Epoch 530 loss 0.22134639755421673 correct 50
+Epoch 540 loss 0.21326376486729373 correct 50
+Epoch 550 loss 0.20564983951648994 correct 50
+Epoch 560 loss 0.19846884799540454 correct 50
+Epoch 570 loss 0.19168832122143856 correct 50
+Epoch 580 loss 0.1852787427037586 correct 50
+Epoch 590 loss 0.1792132385019649 correct 50
+Epoch 600 loss 0.1734673035708013 correct 50
+Epoch 610 loss 0.16801855983232233 correct 50
+Epoch 620 loss 0.16284654195522322 correct 50
+Epoch 630 loss 0.15793250736690068 correct 50
+Epoch 640 loss 0.15325926749207958 correct 50
+Epoch 650 loss 0.14881103761299838 correct 50
+Epoch 660 loss 0.14457330309083513 correct 50
+Epoch 670 loss 0.14053269998435156 correct 50
+Epoch 680 loss 0.13667690835659957 correct 50
+Epoch 690 loss 0.13299455678041916 correct 50
+Epoch 700 loss 0.1294751367428776 correct 50
+Epoch 710 loss 0.12610892581277108 correct 50
+Epoch 720 loss 0.12288691857692009 correct 50
+Epoch 730 loss 0.11980076447385773 correct 50
+Epoch 740 loss 0.11684271175994544 correct 50
+Epoch 750 loss 0.11400555693557328 correct 50
+Epoch 760 loss 0.11128259903951022 correct 50
+Epoch 770 loss 0.10866759828968314 correct 50
+Epoch 780 loss 0.10615473860975419 correct 50
+Epoch 790 loss 0.10373859363438033 correct 50
+Epoch 800 loss 0.10141409583270015 correct 50
+
+
+## Split
+**Config:** `PTS=50, HIDDEN=6, RATE=0.5, EPOCHS=800`
+
+**Logs:**
+
+Epoch 10 loss 33.75235919418314 correct 27
+Epoch 20 loss 32.85948268142515 correct 30
+Epoch 30 loss 31.5066623771144 correct 42
+Epoch 40 loss 29.65750160610111 correct 39
+Epoch 50 loss 27.80131987403149 correct 38
+Epoch 60 loss 26.167785037919668 correct 38
+Epoch 70 loss 24.66200397481523 correct 38
+Epoch 80 loss 22.884712684200203 correct 39
+Epoch 90 loss 20.657544351214334 correct 41
+Epoch 100 loss 19.112345931137142 correct 41
+Epoch 110 loss 19.459879663637146 correct 40
+Epoch 120 loss 17.83294384764803 correct 40
+Epoch 130 loss 16.679996723418938 correct 40
+Epoch 140 loss 15.719071510137669 correct 40
+Epoch 150 loss 15.703546633170133 correct 40
+Epoch 160 loss 14.804445623923408 correct 40
+Epoch 170 loss 13.066678291825822 correct 40
+Epoch 180 loss 11.403197834137822 correct 43
+Epoch 190 loss 9.605408576185999 correct 44
+Epoch 200 loss 7.285074421009864 correct 48
+Epoch 210 loss 4.304000386495314 correct 49
+Epoch 220 loss 3.4198006693554395 correct 50
+Epoch 230 loss 3.0136028308413305 correct 50
+Epoch 240 loss 2.711368668392037 correct 50
+Epoch 250 loss 2.467331442167649 correct 50
+Epoch 260 loss 2.263274347933186 correct 50
+Epoch 270 loss 2.0900745732594412 correct 50
+Epoch 280 loss 1.9396301277348078 correct 50
+Epoch 290 loss 1.8091389841229173 correct 50
+Epoch 300 loss 1.6932449867147714 correct 50
+Epoch 310 loss 1.5903559022792013 correct 50
+Epoch 320 loss 1.497552606799441 correct 50
+Epoch 330 loss 1.4144004570703967 correct 50
+Epoch 340 loss 1.3385405082291562 correct 50
+Epoch 350 loss 1.2697716062501032 correct 50
+Epoch 360 loss 1.2069461436387507 correct 50
+Epoch 370 loss 1.1493443884652235 correct 50
+Epoch 380 loss 1.0967117952245755 correct 50
+Epoch 390 loss 1.0476488254983098 correct 50
+Epoch 400 loss 1.0020368204487586 correct 50
+Epoch 410 loss 0.9598708689621344 correct 50
+Epoch 420 loss 0.9206122736420168 correct 50
+Epoch 430 loss 0.8839076650282672 correct 50
+Epoch 440 loss 0.8495254396937801 correct 50
+Epoch 450 loss 0.8172635150111653 correct 50
+Epoch 460 loss 0.7869203350795496 correct 50
+Epoch 470 loss 0.7583766147677934 correct 50
+Epoch 480 loss 0.7314697635123658 correct 50
+Epoch 490 loss 0.7061127014776227 correct 50
+Epoch 500 loss 0.6821342136253884 correct 50
+Epoch 510 loss 0.6594411318814467 correct 50
+Epoch 520 loss 0.6379700071761089 correct 50
+Epoch 530 loss 0.6175500667433639 correct 50
+Epoch 540 loss 0.5981613319799653 correct 50
+Epoch 550 loss 0.5797782059585791 correct 50
+Epoch 560 loss 0.5622985183204681 correct 50
+Epoch 570 loss 0.5456702017959787 correct 50
+Epoch 580 loss 0.5297827226238305 correct 50
+Epoch 590 loss 0.5146217885303698 correct 50
+Epoch 600 loss 0.5001534644463455 correct 50
+Epoch 610 loss 0.4863190738562789 correct 50
+Epoch 620 loss 0.4730769346467114 correct 50
+Epoch 630 loss 0.46042134263569323 correct 50
+Epoch 640 loss 0.44827767350160963 correct 50
+Epoch 650 loss 0.4366712480135445 correct 50
+Epoch 660 loss 0.42552859095855566 correct 50
+Epoch 670 loss 0.4148430557641768 correct 50
+Epoch 680 loss 0.404564552057625 correct 50
+Epoch 690 loss 0.3947008701407381 correct 50
+Epoch 700 loss 0.3852010314076847 correct 50
+Epoch 710 loss 0.37607071986325724 correct 50
+Epoch 720 loss 0.36728463323116706 correct 50
+Epoch 730 loss 0.3588245684683126 correct 50
+Epoch 740 loss 0.3506886629139211 correct 50
+Epoch 750 loss 0.34284684715298863 correct 50
+Epoch 760 loss 0.33527444467360507 correct 50
+Epoch 770 loss 0.3279736730074926 correct 50
+Epoch 780 loss 0.32092303189405386 correct 50
+Epoch 790 loss 0.31410711315851425 correct 50
+Epoch 800 loss 0.30752631825797144 correct 50
+
+
+## XOR
+**Config:** `PTS=50, HIDDEN=10, RATE=0.5, EPOCHS=1000`
+
+**Logs:**
+
+Epoch 10 loss 31.710840814280708 correct 37
+Epoch 20 loss 27.619434727361654 correct 44
+Epoch 30 loss 35.15578215814027 correct 27
+Epoch 40 loss 19.65814036712436 correct 43
+Epoch 50 loss 19.86269657424437 correct 41
+Epoch 60 loss 15.914213912439719 correct 44
+Epoch 70 loss 14.545861680601343 correct 45
+Epoch 80 loss 14.912667936081437 correct 45
+Epoch 90 loss 15.580266441239873 correct 44
+Epoch 100 loss 13.417874983848856 correct 45
+Epoch 110 loss 12.252437753240343 correct 45
+Epoch 120 loss 11.6841781074567 correct 45
+Epoch 130 loss 10.916858638621648 correct 46
+Epoch 140 loss 10.28385039442434 correct 47
+Epoch 150 loss 9.707644708218638 correct 47
+Epoch 160 loss 9.235796909717186 correct 48
+Epoch 170 loss 8.772067053248017 correct 48
+Epoch 180 loss 8.139314506793943 correct 48
+Epoch 190 loss 7.460210138703441 correct 48
+Epoch 200 loss 6.927219607486519 correct 48
+Epoch 210 loss 6.376593067036393 correct 48
+Epoch 220 loss 5.913140927983057 correct 49
+Epoch 230 loss 5.483373888652583 correct 49
+Epoch 240 loss 5.061634295028014 correct 49
+Epoch 250 loss 4.682677816397556 correct 49
+Epoch 260 loss 4.335117380899207 correct 49
+Epoch 270 loss 3.99520944553466 correct 49
+Epoch 280 loss 3.683357234181038 correct 49
+Epoch 290 loss 3.416203378252851 correct 49
+Epoch 300 loss 3.1486328702689423 correct 49
+Epoch 310 loss 2.9083076827214844 correct 49
+Epoch 320 loss 2.6780420929150224 correct 49
+Epoch 330 loss 2.487639668979288 correct 49
+Epoch 340 loss 2.3241246076104614 correct 49
+Epoch 350 loss 2.1616183847637296 correct 50
+Epoch 360 loss 2.0043098043445395 correct 50
+Epoch 370 loss 1.880820235858625 correct 50
+Epoch 380 loss 1.7791692991943602 correct 50
+Epoch 390 loss 1.6651766645908703 correct 50
+Epoch 400 loss 1.557542048850534 correct 50
+Epoch 410 loss 1.4744329886289733 correct 50
+Epoch 420 loss 1.385967388946977 correct 50
+Epoch 430 loss 1.3124665776994984 correct 50
+Epoch 440 loss 1.2241953077783376 correct 50
+Epoch 450 loss 1.163476776544279 correct 50
+Epoch 460 loss 1.1045837738465025 correct 50
+Epoch 470 loss 1.0523798442926309 correct 50
+Epoch 480 loss 1.0028078975391308 correct 50
+Epoch 490 loss 0.9589416345016335 correct 50
+Epoch 500 loss 0.9115663638093799 correct 50
+Epoch 510 loss 0.8741984577090567 correct 50
+Epoch 520 loss 0.8307487866205087 correct 50
+Epoch 530 loss 0.8015547549050164 correct 50
+Epoch 540 loss 0.7668909377698404 correct 50
+Epoch 550 loss 0.7407965319031712 correct 50
+Epoch 560 loss 0.7137931412422537 correct 50
+Epoch 570 loss 0.6839435648496994 correct 50
+Epoch 580 loss 0.6620958530429077 correct 50
+Epoch 590 loss 0.6424177379021753 correct 50
+Epoch 600 loss 0.6222351333348232 correct 50
+Epoch 610 loss 0.5984583263793786 correct 50
+Epoch 620 loss 0.5755716347010625 correct 50
+Epoch 630 loss 0.5694832067543669 correct 50
+Epoch 640 loss 0.5463241229098806 correct 50
+Epoch 650 loss 0.5332457454154678 correct 50
+Epoch 660 loss 0.5172149217294235 correct 50
+Epoch 670 loss 0.4966068054997729 correct 50
+Epoch 680 loss 0.48983871229031295 correct 50
+Epoch 690 loss 0.468972200777932 correct 50
+Epoch 700 loss 0.45349750377209086 correct 50
+Epoch 710 loss 0.44426382343021786 correct 50
+Epoch 720 loss 0.4352274800525545 correct 50
+Epoch 730 loss 0.4266537576558136 correct 50
+Epoch 740 loss 0.4178681085078735 correct 50
+Epoch 750 loss 0.4010334948330733 correct 50
+Epoch 760 loss 0.3944971270018335 correct 50
+Epoch 770 loss 0.38194219976990984 correct 50
+Epoch 780 loss 0.37775906698736694 correct 50
+Epoch 790 loss 0.3663828777720555 correct 50
+Epoch 800 loss 0.3606369644047365 correct 50
+Epoch 810 loss 0.345528962682325 correct 50
+Epoch 820 loss 0.3403264929221114 correct 50
+Epoch 830 loss 0.3379484035700957 correct 50
+Epoch 840 loss 0.3235983903768161 correct 50
+Epoch 850 loss 0.320870540554103 correct 50
+Epoch 860 loss 0.3172260906216512 correct 50
+Epoch 870 loss 0.3052846719025846 correct 50
+Epoch 880 loss 0.30153148360617515 correct 50
+Epoch 890 loss 0.296975810407801 correct 50
+Epoch 900 loss 0.29088183227443853 correct 50
+Epoch 910 loss 0.28702131065022113 correct 50
+Epoch 920 loss 0.28158399140948914 correct 50
+Epoch 930 loss 0.2799092483854836 correct 50
+Epoch 940 loss 0.2767137545306454 correct 50
+Epoch 950 loss 0.268825390007598 correct 50
+Epoch 960 loss 0.25776259786465877 correct 50
+Epoch 970 loss 0.26017808392686775 correct 50
+Epoch 980 loss 0.25749061311077026 correct 50
+Epoch 990 loss 0.25034927599141327 correct 50
+Epoch 1000 loss 0.24044277636451425 correct 50
+
+
+## Circle
+**Config:** `PTS=50, HIDDEN=12, RATE=0.5, EPOCHS=1200`
+
+**Logs:**
+
+Epoch 10 loss 29.873557603025002 correct 34
+Epoch 20 loss 28.637278694749938 correct 34
+Epoch 30 loss 26.693378698986294 correct 34
+Epoch 40 loss 22.641189507191495 correct 38
+Epoch 50 loss 19.450614781242184 correct 42
+Epoch 60 loss 22.21288506941197 correct 36
+Epoch 70 loss 20.02815629412635 correct 39
+Epoch 80 loss 18.485933278115393 correct 41
+Epoch 90 loss 18.938354344463555 correct 40
+Epoch 100 loss 15.801384177517944 correct 42
+Epoch 110 loss 20.54504173116277 correct 38
+Epoch 120 loss 12.0860808864857 correct 47
+Epoch 130 loss 9.89058856137845 correct 46
+Epoch 140 loss 9.747368374017055 correct 47
+Epoch 150 loss 17.02054927862798 correct 41
+Epoch 160 loss 8.417497986638514 correct 48
+Epoch 170 loss 7.135484223361151 correct 49
+Epoch 180 loss 6.527563031251834 correct 49
+Epoch 190 loss 6.861019469954264 correct 47
+Epoch 200 loss 10.417714715506913 correct 45
+Epoch 210 loss 7.005945422699021 correct 46
+Epoch 220 loss 6.260572947445757 correct 46
+Epoch 230 loss 7.410931169681428 correct 45
+Epoch 240 loss 5.539103577833562 correct 46
+Epoch 250 loss 6.62250431049139 correct 46
+Epoch 260 loss 5.598421674830261 correct 46
+Epoch 270 loss 7.864933680599687 correct 46
+Epoch 280 loss 4.997206273167954 correct 47
+Epoch 290 loss 7.574590575558654 correct 45
+Epoch 300 loss 6.562601431616851 correct 45
+Epoch 310 loss 4.87326656356385 correct 46
+Epoch 320 loss 2.6038494048825527 correct 50
+Epoch 330 loss 2.315534036694687 correct 50
+Epoch 340 loss 8.82038939619125 correct 46
+Epoch 350 loss 2.9707709095829307 correct 50
+Epoch 360 loss 2.0873986412435075 correct 50
+Epoch 370 loss 1.8453401831690723 correct 50
+Epoch 380 loss 1.7092717007677 correct 50
+Epoch 390 loss 1.6084180871732447 correct 50
+Epoch 400 loss 1.5280966883231553 correct 50
+Epoch 410 loss 1.3874388714658201 correct 50
+Epoch 420 loss 1.3605726791706867 correct 50
+Epoch 430 loss 1.2963766574678814 correct 50
+Epoch 440 loss 1.2152162535923534 correct 50
+Epoch 450 loss 1.1423989193296675 correct 50
+Epoch 460 loss 1.0568357842479614 correct 50
+Epoch 470 loss 1.019782761902968 correct 50
+Epoch 480 loss 0.9570502900552107 correct 50
+Epoch 490 loss 0.9173540575276944 correct 50
+Epoch 500 loss 0.8703217240424889 correct 50
+Epoch 510 loss 0.836076807034688 correct 50
+Epoch 520 loss 0.7989480024734267 correct 50
+Epoch 530 loss 0.7616411283425616 correct 50
+Epoch 540 loss 0.7322332364904421 correct 50
+Epoch 550 loss 0.7026049137321023 correct 50
+Epoch 560 loss 0.6734059435788355 correct 50
+Epoch 570 loss 0.6514508905857581 correct 50
+Epoch 580 loss 0.6253657491441774 correct 50
+Epoch 590 loss 0.6016419327474636 correct 50
+Epoch 600 loss 0.5808687869112696 correct 50
+Epoch 610 loss 0.5612780091994775 correct 50
+Epoch 620 loss 0.5418942382173502 correct 50
+Epoch 630 loss 0.5236758362305461 correct 50
+Epoch 640 loss 0.5075390663097562 correct 50
+Epoch 650 loss 0.4911208096803893 correct 50
+Epoch 660 loss 0.476981659802743 correct 50
+Epoch 670 loss 0.46137354360866134 correct 50
+Epoch 680 loss 0.44798499696600164 correct 50
+Epoch 690 loss 0.43528270400602687 correct 50
+Epoch 700 loss 0.4222083739843732 correct 50
+Epoch 710 loss 0.4104886183463712 correct 50
+Epoch 720 loss 0.39895653616710464 correct 50
+Epoch 730 loss 0.38878342428216206 correct 50
+Epoch 740 loss 0.37886167579833563 correct 50
+Epoch 750 loss 0.368570697147489 correct 50
+Epoch 760 loss 0.3589496718668607 correct 50
+Epoch 770 loss 0.3499901651590006 correct 50
+Epoch 780 loss 0.3413641537631822 correct 50
+Epoch 790 loss 0.3332250243507555 correct 50
+Epoch 800 loss 0.3251315015013947 correct 50
+Epoch 810 loss 0.3178095612824394 correct 50
+Epoch 820 loss 0.31041912002011607 correct 50
+Epoch 830 loss 0.30319841873983694 correct 50
+Epoch 840 loss 0.2964133607017596 correct 50
+Epoch 850 loss 0.29008566506978545 correct 50
+Epoch 860 loss 0.2836831216252566 correct 50
+Epoch 870 loss 0.27764790133186296 correct 50
+Epoch 880 loss 0.2718419682640759 correct 50
+Epoch 890 loss 0.26620205028629884 correct 50
+Epoch 900 loss 0.26070236509732264 correct 50
+Epoch 910 loss 0.2555126126082473 correct 50
+Epoch 920 loss 0.25049285536606997 correct 50
+Epoch 930 loss 0.2454331491339338 correct 50
+Epoch 940 loss 0.2407241228091517 correct 50
+Epoch 950 loss 0.23600127713997474 correct 50
+Epoch 960 loss 0.2315757283112603 correct 50
+Epoch 970 loss 0.22730738692423827 correct 50
+Epoch 980 loss 0.22314738218284338 correct 50
+Epoch 990 loss 0.21922294110060087 correct 50
+Epoch 1000 loss 0.21508616884109666 correct 50
+Epoch 1010 loss 0.21131815650591748 correct 50
+Epoch 1020 loss 0.20752792859416885 correct 50
+Epoch 1030 loss 0.20404733796003943 correct 50
+Epoch 1040 loss 0.20054067812529294 correct 50
+Epoch 1050 loss 0.19711702710764678 correct 50
+Epoch 1060 loss 0.19374560749859682 correct 50
+Epoch 1070 loss 0.19050864188466443 correct 50
+Epoch 1080 loss 0.18745335981221015 correct 50
+Epoch 1090 loss 0.18441274555208223 correct 50
+Epoch 1100 loss 0.18148126665642772 correct 50
+Epoch 1110 loss 0.17869402481144853 correct 50
+Epoch 1120 loss 0.1758290920335748 correct 50
+Epoch 1130 loss 0.17310900893611622 correct 50
+Epoch 1140 loss 0.17041855936055722 correct 50
+Epoch 1150 loss 0.16786323307966283 correct 50
+Epoch 1160 loss 0.16529510067226796 correct 50
+Epoch 1170 loss 0.16290304693243238 correct 50
+Epoch 1180 loss 0.16048035852185277 correct 50
+Epoch 1190 loss 0.15815382883096868 correct 50
+Epoch 1200 loss 0.15586020629583122 correct 50
+
+
+## Spiral
+**Config:** `PTS=100, HIDDEN=20, RATE=0.5, EPOCHS=1500`
+
+**Logs:**
+
+Epoch 10 loss 68.71511608037055 correct 61
+Epoch 20 loss 68.30588930020942 correct 65
+Epoch 30 loss 68.00332527042707 correct 61
+Epoch 40 loss 67.70521657149985 correct 58
+Epoch 50 loss 67.47138855991601 correct 58
+Epoch 60 loss 67.26976357327436 correct 59
+Epoch 70 loss 67.12410826307537 correct 58
+Epoch 80 loss 67.00707320601778 correct 58
+Epoch 90 loss 66.91497041301285 correct 58
+Epoch 100 loss 66.83572070894556 correct 59
+Epoch 110 loss 66.76864379993027 correct 59
+Epoch 120 loss 66.7022574516325 correct 59
+Epoch 130 loss 66.62232360866038 correct 59
+Epoch 140 loss 66.54962864111374 correct 59
+Epoch 150 loss 66.48262495947542 correct 58
+Epoch 160 loss 66.41083748439603 correct 59
+Epoch 170 loss 66.3664138202179 correct 59
+Epoch 180 loss 66.31367467459017 correct 59
+Epoch 190 loss 66.2907758410631 correct 63
+Epoch 200 loss 66.98089399261403 correct 52
+Epoch 210 loss 67.06177733008565 correct 51
+Epoch 220 loss 66.33192033937969 correct 60
+Epoch 230 loss 66.15057941730524 correct 61
+Epoch 240 loss 66.10070291524839 correct 61
+Epoch 250 loss 66.01192857278852 correct 61
+Epoch 260 loss 65.97473727265219 correct 61
+Epoch 270 loss 65.92903057040971 correct 60
+Epoch 280 loss 65.89971517954437 correct 59
+Epoch 290 loss 65.89941068415979 correct 57
+Epoch 300 loss 65.87273454966903 correct 56
+Epoch 310 loss 65.79172252179588 correct 56
+Epoch 320 loss 65.61825667056381 correct 59
+Epoch 330 loss 65.49223403247998 correct 61
+Epoch 340 loss 65.45810445493414 correct 59
+Epoch 350 loss 65.4260714391121 correct 58
+Epoch 360 loss 65.42439498679903 correct 57
+Epoch 370 loss 65.12558840760677 correct 59
+Epoch 380 loss 65.12263175748602 correct 59
+Epoch 390 loss 65.14940130288834 correct 58
+Epoch 400 loss 65.00537990021363 correct 57
+Epoch 410 loss 64.89748060567388 correct 57
+Epoch 420 loss 64.7885364700252 correct 57
+Epoch 430 loss 64.84280871319638 correct 54
+Epoch 440 loss 64.81893998341585 correct 53
+Epoch 450 loss 64.61214677100281 correct 54
+Epoch 460 loss 64.39150353735262 correct 55
+Epoch 470 loss 64.23123804962226 correct 59
+Epoch 480 loss 64.29466188549844 correct 56
+Epoch 490 loss 64.47516768150778 correct 51
+Epoch 500 loss 64.48738324125286 correct 53
+Epoch 510 loss 64.32293607711219 correct 51
+Epoch 520 loss 64.00080281055355 correct 56
+Epoch 530 loss 63.961607833575904 correct 55
+Epoch 540 loss 63.93100987939193 correct 52
+Epoch 550 loss 63.82236663127581 correct 55
+Epoch 560 loss 63.74195256644413 correct 52
+Epoch 570 loss 63.64833957025323 correct 52
+Epoch 580 loss 63.48868699116711 correct 52
+Epoch 590 loss 63.3862750280804 correct 51
+Epoch 600 loss 63.28512438287466 correct 51
+Epoch 610 loss 63.22309156486676 correct 51
+Epoch 620 loss 63.113414011986976 correct 55
+Epoch 630 loss 62.97173427564117 correct 53
+Epoch 640 loss 62.986479652566075 correct 51
+Epoch 650 loss 63.09358328696164 correct 52
+Epoch 660 loss 63.12407676671131 correct 53
+Epoch 670 loss 62.948836110456575 correct 53
+Epoch 680 loss 62.83578457174591 correct 53
+Epoch 690 loss 62.69799755893202 correct 52
+Epoch 700 loss 62.74274934766089 correct 55
+Epoch 710 loss 62.682703581533374 correct 54
+Epoch 720 loss 62.58973462419676 correct 54
+Epoch 730 loss 62.29942212185198 correct 54
+Epoch 740 loss 62.256309115529454 correct 55
+Epoch 750 loss 62.67994497845996 correct 57
+Epoch 760 loss 62.35083030795144 correct 55
+Epoch 770 loss 62.25955261906933 correct 56
+Epoch 780 loss 62.15185472664485 correct 58
+Epoch 790 loss 61.83685263176922 correct 55
+Epoch 800 loss 61.66471757719838 correct 56
+Epoch 810 loss 62.56581629317485 correct 62
+Epoch 820 loss 61.89923803491597 correct 59
+Epoch 830 loss 61.53355410830513 correct 58
+Epoch 840 loss 61.80143358072765 correct 62
+Epoch 850 loss 61.82515412052376 correct 58
+Epoch 860 loss 61.68465874367569 correct 60
+Epoch 870 loss 61.281307286153734 correct 59
+Epoch 880 loss 60.74353654336668 correct 56
+Epoch 890 loss 61.276936859656935 correct 61
+Epoch 900 loss 61.23453721339674 correct 60
+Epoch 910 loss 60.68750857246043 correct 60
+Epoch 920 loss 60.52102737524052 correct 60
+Epoch 930 loss 60.92190154409212 correct 61
+Epoch 940 loss 60.80069844809939 correct 64
+Epoch 950 loss 60.551295478144404 correct 61
+Epoch 960 loss 61.35047329759586 correct 62
+Epoch 970 loss 61.201626802506546 correct 61
+Epoch 980 loss 60.276827760684796 correct 62
+Epoch 990 loss 59.539994015615534 correct 64
+Epoch 1000 loss 60.80067748852008 correct 61
+Epoch 1010 loss 61.67082447175194 correct 61
+Epoch 1020 loss 60.43011429772951 correct 62
+Epoch 1030 loss 60.24125283261178 correct 61
+Epoch 1040 loss 60.216744151713804 correct 64
+Epoch 1050 loss 59.89960620304444 correct 63
+Epoch 1060 loss 60.3740579607568 correct 63
+Epoch 1070 loss 60.07679982710797 correct 64
+Epoch 1080 loss 60.00566934436829 correct 65
+Epoch 1090 loss 59.70365346169148 correct 62
+Epoch 1100 loss 59.91358188159955 correct 64
+Epoch 1110 loss 59.473717162404114 correct 65
+Epoch 1120 loss 59.78432820756897 correct 62
+Epoch 1130 loss 59.47552114117245 correct 68
+Epoch 1140 loss 59.10096301256793 correct 63
+Epoch 1150 loss 59.630560045185945 correct 65
+Epoch 1160 loss 59.28043412404348 correct 67
+Epoch 1170 loss 59.073255536559856 correct 62
+Epoch 1180 loss 59.16488798942276 correct 62
+Epoch 1190 loss 58.8962009236161 correct 66
+Epoch 1200 loss 58.56811011623168 correct 66
+Epoch 1210 loss 58.86981297452959 correct 66
+Epoch 1220 loss 58.93937821166227 correct 67
+Epoch 1230 loss 58.536200435628366 correct 66
+Epoch 1240 loss 58.91699684381775 correct 68
+Epoch 1250 loss 58.85785353833471 correct 66
+Epoch 1260 loss 58.543703602179086 correct 66
+Epoch 1270 loss 58.030444406997006 correct 66
+Epoch 1290 loss 57.81490957159442 correct 67
+Epoch 1300 loss 58.00140726690014 correct 66
+Epoch 1310 loss 57.934715432543896 correct 66
+Epoch 1320 loss 57.99325114718681 correct 65
+Epoch 1330 loss 58.607947056997354 correct 67
+Epoch 1340 loss 56.50163935605804 correct 68
+Epoch 1350 loss 57.565455468309366 correct 65
+Epoch 1360 loss 58.327873310157486 correct 63
+Epoch 1370 loss 56.52132160696343 correct 70
+Epoch 1380 loss 55.87205126197996 correct 68
+Epoch 1390 loss 57.31080309061276 correct 68
+Epoch 1400 loss 57.35315618763012 correct 67
+Epoch 1410 loss 59.84551149331509 correct 65
+Epoch 1420 loss 56.448622670255965 correct 63
+Epoch 1430 loss 59.42861474237802 correct 58
+Epoch 1440 loss 55.098119209084246 correct 70
+Epoch 1450 loss 53.33741456176819 correct 70
+Epoch 1460 loss 57.80002625049885 correct 66
+Epoch 1470 loss 61.88064947510794 correct 59
+Epoch 1480 loss 55.72888569952491 correct 66
+Epoch 1490 loss 58.11701591073116 correct 59
+Epoch 1500 loss 56.503936394488356 correct 63
\ No newline at end of file
diff --git a/logs/circle.txt b/logs/circle.txt
new file mode 100644
index 0000000000000000000000000000000000000000..0fb91ea07b0d6ff85e6146c387447712ab04a8d4
GIT binary patch
literal 11788
zcmb8!YmOsD5Cq`=H1S8LkAL9dX5`@GLBoJ_ZpuP^8)lO%kDp|IByzP8OL}q#a
z{_BVBx9!*M$L;6s%XqwQm;L>h@!22a;j_2xxxH*3+hcp*ZsYOT?&Gfa@q8P9UH11c
z!@{rI@8j{)xcYN^>wUX!-*4Y--`am0ceDAe{f}Y)KECl779YdYeR#Z%tLu1wj?X{F
zH|Fe%>TLhM+h>W7*#8)oUCHh`ZU0)@f7|U}NA+HZ&F5Z~$GCouryOpj{g0wP>iIc-
zc3y@Z|90Ok%HefHejQI9tN8ufx&Au$+5X$O^JTa6Hf(+T>qk}O^Bfhcs(&x6KT7MD
zVI5Y)eEYBB;@9)KHqXNPM`2y|Y9s3OHwY5>ZCLl&IeM9$Yx8XF2ZLh04tv-Ak=XH}
zvVSS;51YhzZig_85=b{;mu2-Iev(1J1qdu|rVozmGd#_g?UI&OjvR3hCL|
zzE5=+o4k6uPFE`HxLsMV*ca6zaN->Mb=*-|uei%PouO^_VUb<;=>dIOZ}QL8ABcM4R+U!haC&b1HvMay<9-p8H#2SGJ<>fToD6}@VUUA<`q
z71aaCeJ$g!*o&#>_$l*kuZh0uKV+IK?#H#NKDGUdzwE;zUeGewFjM41e`@z=${@40*D-cS{&mMvw=n;M57P%~_YTbc~0eW|~uR`qcVE*y~O?*+I
zn!mOEiofDlpUbd(sLpfvGY#~hW{-}^m+@o^U-(wjTUD*s(Y}M9;|e?;_r1-9}*!4;;EVn$s;VQ
z;;;Et@fUh@6dE%FVA4ZY*ihN8(4(R?`hybPfyhKJgDB{{{ffQnlg(K-MV!6iQ0U5f
zh0XW9zUR%&-VF8dgSn7t`a)}T_3QzT&+`#q&h;_bRfIKa%$Gc(q%xbgU$NKx5^ZvT
zQMkWAlOfoF{0V%{&o7>%51G!LNIT
z;6L@xT!c^A;9WIgx*RGF^N{LbZ#Ey6ad@RZ6@S@QH~7p305g0*v?(+10e;Rs<^j@W
z20M65Kh@>H?Q91!r#`cc2MnV^-je|w21#^d-a|Em<3XKh9i^b2>4bdzJ@297F%K|L
zcXU?l@?lv<^`=;(NZ0Bg-oswvAFazI~5XtjRPmp;nMoF*+W?
zVDmSa&8*2>n6Hrt&t2dF$IJv+!U%qBNSSWc)>T)@IimAml11lKGXCHKA7k^c)R_#r#^bz)iqAaIg-p(zpblP#;q!_SPE|S3QpzAlzy4u_-H{$q*7+PdMjvW=
zs5a%(1KgnE^By`zs~*g-8BJw36GGo-J+M=ES}oKEzi1z=2yR`ic<2}%+|$%BF7qMh
z*_GL#+FcInqsQ=X7l=N=6uiz?spB*~QSV^KmA@g1GRtmn90L;kzOhm9SudonPXK|byjwww+q94(lc?E>>&Ij#-QwYN7
zY$3;QVe{PZ-OdN(WE#S(DrHh&_k1rpZev^&7AIH_@4T*TuQ~Bo^BR9K4r`(xPA&cl
zJLYrfxXlBd$UKE8dgVN-<12{mn60DP%`$ny^eDSse>MjhbbWS$lZEQ2R9cQ>2a^P*9#c@t_M}8g1|6k~j3Alo&j@db(P!IL``YQpP@j2h`Zf=`$mL&y0KF^G3TBU(}h!`IriWgcQ
zO<9?Ew(Jg5l!%7#mRDQUaT1=blwXi~4>>u)7GH9a8Rp4n-bcr8(PjsPr0lj(Ws^12
z=RNf74tjV}CAsJZ#l`V?4?Vl18Z;vuq3fwoX7{|0p55w`Q?Fla!t$f*^P7jRb=?-b
zK6%9m$oP5^Td7pfZhXi_!v~e}igl~~?)hz6WNS7bOrvZbURSpAe6OokCbl}+a-tl*
zVPdT4t?zi&IwJaLEPqiFQwehI?6b(?q33zjRsU5E*HgLVH9wtr>6xAr5=Q7XU(oIw
zyv<{nm!9n`QWPp>zT${;zsy$l1ztMFLzp?G3+n1x4JmuPuJonjJTK}*A>HH3jLKGm
zH!EH`-t&@7VHsb7Wf*pfUh7N8d{&c*-pe4Jp#bo|hOv(O*&L$t@n&kly=u{(`HG3F
zW+g8)EE$CDIz%gNNGpAbI#t|}8s!em;Tk|Hs%rpEuCr
literal 0
HcmV?d00001
diff --git a/logs/diag.txt b/logs/diag.txt
new file mode 100644
index 0000000000000000000000000000000000000000..7d983a431780fa6f99427da1e79577f15b8a831b
GIT binary patch
literal 7862
zcmb8zTdpHj3-^|77y$3N=5zw63-kL{Q3TAzJx_xJ?hoD{;u1z-uY7B|F->A
z*KhUgpL%_*>*wvK?fdBGW%TnN|5CB8``@h|AC=*>-RhdgzbxaQDyq{zEAFM@p6g#H
zSMPNk|Jug?EOb8Ww9fvl9CF=vv_AIN_i_B&GQJbv_j8}>@lu)Kblx*f(Qz_-e$VZ_
zPLD}wKkK^f_`pUEnWphS+xXXtA55-$1>%beB{`?jA8qvWetNp0pv#W>XI1f07j&l4
zpH1{<#im0_#0+$}z)A4i>+;^{p32^_r<>c3iFh(NHHXRJlkW6={LZ(S4G!?fF`PK+
zRL{Kb;y>E+ch0?4e09O+T2bMHTaVC^X`Z3;F6}^s_T_*Ok6r2&etY%w6?ERc>i5+bOL?nUeCft$M7mc=U(c^D9u2F3Lx{W=kSv2
zoq<(`&c8A|D^{Z8R^Nd=rXTfs9=~&N7)Zo)W;*yR%+#cs_xKdhanN}<9B|XiQ{w9r
zFCD$d@A2V)caGtsAHyS^Q)l32sCj?RK=}KyL%%v4eVQwjn(oPWkn6hKs)o+TAG?ZZ
z#;25|63-xty-8GT=EgfRc0zbHO{=yyh7?Kg*D5
z;MBSBKvCwu?mExVIa&srgGr2O44BboKIPKbG=sX|nN9|A^$ARgnF-J({v%(0+WzRb
zWYG08Jz)hgEYp$fbCfzqCxcTvJ*{zp4i%5XYM!BUv<#UuX1*N>AG)6DX&X3#(#p{|
zz-{*RxLPMv1q~djV&!R!Vm6zAZYnFiW;>hXwDL4RtHCanO^Lf-dR5;vhi~_;d+jf<
zkB(k}9!
z(=RLIPhlb*bqcR;RMv2B0k0e7{TJ@G|UJFj8o?aZFs^JlJx3pw#|
zgo%~A^$4+WK@aJ!*^Rw_#1|`fQ&Q#*#Zr{1ferZ0XN6wZbCC|K3h`-zA80C@`IHM>
zuQJ$&gH`&T{vi|B=~wd{D{oVe3NyRYX*th(@S7v$TdO+6)p_Sgj+Mhp$D~H>nupu56;FY^!*s5`!V>Q9$NPMj88(3Dg%v3ecv2r*M@SP1R9Ts=#keE4=
zW9{(l<}iw?^6RDerqdD5*A9o`>lY#>GWsh9eG-K6{6UcBPOm;SCpaYcmEz~!d5*Qq^>EsS
zX8K8(@eK-_=UBTu(@DI{Sg57*s%5TR-d(YQRa!MMawii-W%|wMA3oPP}tky1fcCoUHQuu3QaKACXRqmX=Po&1t2jGc
oTsX&S?eg^6CTDxEnd{0dU1J3if6mW?xK3YC-@1n+a
literal 0
HcmV?d00001
diff --git a/logs/simple.txt b/logs/simple.txt
new file mode 100644
index 0000000000000000000000000000000000000000..ee4031ce06bd5c97f61fe2d733efc8b41c2d1a4d
GIT binary patch
literal 8456
zcmeI1Yl;(55QY1<;2vDTkbb4R7vM?;!5@O6u3tUhJqC4B1=Wy*NswlebdssM^*D9T
zz4`g$efOpN+&y<6yLYAi2dcyH{6?y1{!P2H?-9Cbf+OS^h~+Uq;d`uZibdFN;rlpU=h?!y@bUgFkW?tv|BQddTp;5V_kE$zQ1KVS>
zO%)Hi@J?5f_fgN2?yA*(HL$T}#=%SNO83co<}DZ1p79a2)|_M(DFj=dBwaW0^<%m*JHT8hiFbPL76>t&7}1c0kKeb_tdr{
zhGRh9i9Nma2r)!ck1SCIJ_C84P-Y%bSD)8#>p7xrpU!0Nk#~HZ
z7@11!nS1vh0UaLah`xuBiO^_P{`zBx7?e7oNWn9)SU15zRvxj2kCiZLD9#eQ0d`PK@RPEHKbf#pr9Qps
zn9RpLMUg3iKYJW>Oz&~n_
z8Df>T>8`3>`Jb#3RDszI8oY<213cO5y3e?cc;bFyu$qWtfTh?~`7D5L&p_PuH)+
zEuy%EkBqZX(LPBuXMk;=WWj!WW?-f}u)`h>)yMuCMaae6gk2|n;eIKAZn;<@UB$rkLj)*9TrHYJOc*r;pt|jw6`#!Co6@h#7R6UAY_{g*Y
znwZt|N3YVcM8TewIO=^1Q^xhaq>qeSnX*LT)l&tO6l=2Psq;6Cmc?;7ALKaFgT!GD
z2iY(MDDl|=_jQajeXj2#<0i5WHb9AgMW8yTi>T`O$e05kFe{Uf$xGGQZ_saa@1EYx
zm;;tLO>!nYf7A%pu)RCeSnT^UiqxkdI4Y7Ri2JuGWYQdKe;QE
literal 0
HcmV?d00001
diff --git a/logs/spiral.txt b/logs/spiral.txt
new file mode 100644
index 0000000000000000000000000000000000000000..fc015803ac1722c88681605a1e7bb9bcb69380c5
GIT binary patch
literal 14524
zcmb8#TW+hz5d>g98^|8K0z(&4A{XG57y<)4>dF27!WT#whw%klh=_1bUi;kA#;{qoc0bN%kNzFybQFYA?;%lmrm
zWxaM=@A+lD|KIEJ_x0WH$7fzIf4lr}`Lp%6(fa#g>Avhfmfe^2@ZQJex}KZ$Te0qQ
z;@x9s?|nVqR;+JF)a;7>vs!PlU#s=c0~sj19mw34#n%<}V?FK_ncI4ON9HzKzaAFf
z*QeOK9hOsr*AjcTu3|efVt-rJd@K(i3mvh;=yS#YzFygJ=Y5YIj@|m}vUXi~=og+;
z=yUneIn8=QMyGY_s+t;#`eoUFS-w8k?`)qfjM6WZL4E!Fy!Fml=zJWu
z+0%*8`8b~O;CcIZ|DW=zVO_TA|`IJ3;h56&CGtHl{f2;O8b7?Xqi;mG_9N~>FbZ)h2g_%?QY`z|G
zXDZSKl{E==^s%quGjrFTU?Q>!ds$>cHg3ta*`K)^Rn@!d%o~=`(;k1fKXX?f1|O(X
zXcNWveHZEW`^>$q`kJ@Y0P1YhiR=&GiNA9<_S64C-TPq4Yw!PF-_Bn4O{!R<`l?LV
zGH3Sw={rLr+_mk?{>j?#mz7CJav;2
z_1V?!cmA@EF*KPXP*2>VV)t6#j(>Ee<6~lp9X>F@E;-S?y}q5l`RrsL7Vy(Gset#>
z(B1wNf9EfJ%vkKpCQcRbSsj}7wsU6Zn4*UVt8lVC1e>XQdUww36{#&yx4_>XXCjzp
zE&jH*>4m2L&|yEjxE{^+CJSo%y!}3VV;?Re&(3UP_E6zwzv1uGQ~Z9D6NHZRlgg@C
z-+UVW(U!fw*M$#tqt6h83U1NriM`!^XD|EbEDS&QJEKiF#fS)MXHSCT;8St6PshYQVqKmkI1S
zn|i~i-7}^JtN_a^(eMfcSK>QNy|+U&R4;u8}-a?4pR=f~`F
z^O?HsKG2=*-w({`0UeFn>DZ1>>mTt=_F1)CM8U;JU)H?vd_A1_eDeX
zrwQZvdHbEg>{A|^1M_rg`n7jY>wnl!)z$S_lO0H{^-vV4Mc?gr{)#@mqcqEES`O7|(&t3C8=cN+#u4T~rQ(rfP-j;`)jLenXF((Ut`~KDxPTsS9@n$2>xiKuT
zDRvF&7|jfZKHZ5%2c4)=@1dp;YSuDH4RlX7Td1GiPb*plIzpYnsX#bp2Ed(i!YE4m
zyob(V9w5V-4c)Wa`cBtAZ@+Vx{oKh@2Q^g*717mwM{gC#X1Q+>?2DE{*aY-xs*pRa
z7JuVFcC&7uy0M=v%bcX39(*)-KgiVVcWh0I5
z^)s+}yYCTlo#M~z;y)8WmHE^~9id)>&f@e?&ZTCfK1#33sw31+8ql0FzyKX12VcCI
z4AUmsFz>MRCGo${BoU-|j(%J4Woc41$rq4=TCEau!pEHUoMM
zu%_9EkhpUWz*_yO;`X7&fysX7F#EKPcSxzMpUpha4WHISlx2#aT1NT91Ge*8x8L#a
zQ&{}bS2e~3+Rwg-ItSX6H>0NbsgLc!duTN^aPG!E-0yz5nH
zb=K#d$^YL;Zk%%Zzya@XTGW+-htB2bwJO0$|D1ClJ9-YS4&myKlm2oufr(B`_Fi2<
zF?3Tq)$}^>F!NdLUpIa_783S~@1Zm^nqEXdc@jCK&fD0=h^f|%)#;~fE--^Z@(rtd
zCEEEFPMy^%k#iCq4GVI%$BUN1WWTeT{qyXN6Jefs_-ebCJ#(LHJ~6Xe=i^!CCv;3{
zzPhru?U#p{)w0Xv$6|fNm)`9tHRa3gz|M!%nKk-Kwe2DL31XC^)uE#lAAeQvc!+L?
z%Td`(hH$X?=s4ljltoT{xCZ^NPMT^x;nZ0Dl?6<25B~nmuF-L7n`%Be<)9PvLhjw7
zAXsivH3$u3#ijn*qagRV(<8la_ILdi=2D>OUc=HUeEs2aGG-HtWGzH
z`{*sX_A~2h21SQab~ZIEG(BX~d`z|I?Dki_Sc%_G)9mu8pQp2L
z-wS-QWuX>qir2KW1Lls8w@ITWq-K7r1C@nM`a;*{l-GIDw~uJVJ}-h4)!+$U;Wym$
z)m{q~Yx{UkIlBjcl_h@tmiq{^RJEhLy9Y>i4>Pxejq2(;@8cyZ^1KJ#G*vqa4t0y9W2nWy=`F-2{(pP;TQ=ZP?|0arZHGEVP@C{2ha+Rm;0%qu#u>
zTFf3gvonqK8dTInBx;_&s@ij)C>L+}#FGp#0)~UO@2VKxT^r@{9Sj|f|
z{G3iz!Dh94m~-g7HYspHhMCogA1f%;?qSLSx;skhnmLgfky#Or&~m@ueRvm~{s(H;
BG&2AI
literal 0
HcmV?d00001
diff --git a/logs/split.txt b/logs/split.txt
new file mode 100644
index 0000000000000000000000000000000000000000..87a4a53882a78f7b191aae62b645a7cee48ad7ca
GIT binary patch
literal 7812
zcmb8zYpUZ`5CmZT4rC9r0^|2%0ke`2$dAB;B2xuQSUxYv3=FZX&C^<39Z^|y_FsW{u{u3GPpYd=-=
zTYYk04;4J>IIS7FmCL)0eQjc&3xz{Py{*_BE8>0414Hpo?aE(GbnE+EIRxP%7L;Fg
z{wP$A-!Xr+(VtCptJ+4tS9B|e2K2>!eD5u+p40IkZT#zs&Are(uCZV1>UB0e>blyk
z9s=9dhcnN@LodS_os-M!rCoi8{i}_C{9k9Ul?CtR!ApARnR9B;$A)bM`U!tAMWZTR
z)@v21Ws4BaZMo@JqYR0RIj`#Jy`uX|hI^Z#Q5|J;2|t
zf=;{GNnh6Pa2B*U54<{#jxux##zoGi8^`rsm!mj!-0OPRg}#PP#qSKzxD_txflgD6
zZtp>ub-|9;Hst7<{O;LrE{>
zpo*wLo<75+gZotn)AS{BdzQiCj%0I|o8_??!;r{G7$3g+puE9u4j`a30yZ<6s-9HIc;
zmu%{pkKgH7e0y$$o_R;&UVR2pu~jeNMGx*tFp7WAqcl=9wP6ou#i7Jb&+%(M
z!${HQ+I!!aT}5<BQc9zJ^ZI$&ft@on>DUuQeIy
zZazb&>14pt%+WkQ5Z9red1s}fzSGuf)!`G3!At+uHb13tew4w!n&x$Z!kzJ{D19EK
zPSa4r9&zJ;rgiRbaZ2v_44tNvf%Cx~SM@*@>Ii#hGeEaTDN2sMqBF7eI4(G!VWeqT
zy!S@(xaY~UDag&cH^&^+pvz{ED)9(=;l@R}yApV8_uQ%;We|~1i;qv)TV@jN$uQD%
ze2>0h1SR$4j(v~fvc^yM#w9z|%;|Wyj98lM^R8jv(4&+cg*~LGd9tZz`mE#gTpDSb
zJILa5k>9gaU8cr-hLNVREYT>A%JL*2cUtpv8oDEt;u+Lt)v&@lT-r0IbzzjjCw6$I
zrupXJPtcpMAqwiaWa^9ee&&-sCZtp}uLr;AYv6NGq2u_qI&_{J&hG?r57pNI!Q6d9
z<~{9_&%SS#O;$Fl0#Z=P?8%(VKG~GplY{s4bzn}AakrH{fK0l+r`sc4!v=5iYzD92
z&f%)u^GJ=9ZBJ4<8!>g%sUi%f
zj(PX<%f5oBEWXDt-tpBe^K16)(0=NKv5
zPD9K*ZOxQ`s_$FZ9!}|9-wW~0RB_WYk?r&M^Y
zF9YFiAl=66G0wVfpOeG;7U_sC{t;d7D)QgzIJMK?epKasJGW}BTQ>Ld?3V)lHbs99
z{B^uvTb#$hg{@YP68)nK-!a5?X11Wudj3~De!Ed;FrUx>wBJhX%cIM>a9CC5`
znxXaiEA_$(M@0ya)Q2LtgF00lx)&V6^XmK}a5K16sJl9yiq-B@`2&wR;`0aUcx?$zw6e+&n~*pfK(rM$<c#jFdJ#7xDJSUwwbbbCxzwp%tSExyaIK*q~
z1TOW@U+LF`fn@^dE@yLYwKCKC=#_e*t4=bY0l0hzmuK!)SEo|%J|Fo#JRH&s*>LMx
zy%j!g9@z;WN8-gW)ljc>0Gs;!q@ZI2^H*kSlqul+aSL(vquT>>yN-aH`s2jWu)Qcc
zFvt24DhY)o&5nzxyHT~(L_IkO3S&=(^FN_
zS-!E~mt}(XBUEb6rkJ^?_Svb;H92J_cO&4@Q41pEPS1S@E#PhTx_Dk6zfv=NywO#R
z)o0#81@yq7`VlHMi-4V~Le)OUi}b_<>f={h4*sq)y?_au-^r)aTCD_CsC!`7FbA+c
zzL(h<2k&bw``~A-$stoX8cqf}uICusuTeFVgO5AeP~gWg)Qj@^TFb^?WnjCa;cQzu
zJJs=64${wIre+GkOmipvj1$>`DXt%(QgiOiR0wnSeyC4AbEj?u-#RjBzxOlActv(z
z`Ph5atzRX(LLGtggsZq8J&2$`6xh$9(z6`)*`FE8FS5_bTJ(N|O3{g+>YNYX{5R)M
zFZzo1&ALkQYrWp9AJ(6?SgDzky-@{YxVdt|I^Q(S@O4Ese`6A?QCoagV
z#5>c6YCemQU6>xQKhG!hV)xg64wa@8A>84trUXWCJ+X*Y!c5b8tm<}*cwIWdrF=TWrkSdBIy-GokG+pE55i=b*9pT%N<^{>K%kI+Mdp*BnKcs9VqDQZg2yKO^K`
zS1C__!zU^T^E_FCe9rvfSgwx6_4uJ`-c6?978G
z0(M^bjaB_hh3gfcMBsjY+GTRzXX3eXCc;eFSmdm^k2XQMRkndM_%l_*$5yyF#Z_tC
zGuQg*66PPD{b1j$)7dO3A+ac%d+_i6UY2jyUC2bsq`(Fd)yW1@JqexVFjIB*3>?ao
z3xl|cSY2P@!?PM?I_pB_f=)ZVDc+w#IbmI+xV0*}cgsD5eJoe{Im{KEJFVP7;6t<_
zGg+Kr>rBzuNof${;20H`>Gv6W-G^)!o-4&q_i1PL<({HjC(j3GG14XdRV)61cf0EP
ze4^BO3>DXfJgZcvOd}`jyKDZ>TE(b5Pf|o3Kjlc9`%%*ibBYTafCh|yLYA2U|jG4BR
zhxvf5;VtIE$J{gBhEelLm$EhCmL7zw;sLt>ye4+|@&ZGPb{ua|8
literal 0
HcmV?d00001
diff --git a/minitorch/autodiff.py b/minitorch/autodiff.py
index 2b69873b..7f90bf24 100644
--- a/minitorch/autodiff.py
+++ b/minitorch/autodiff.py
@@ -22,8 +22,16 @@ def central_difference(f: Any, *vals: Any, arg: int = 0, epsilon: float = 1e-6)
Returns:
An approximation of $f'_i(x_0, \ldots, x_{n-1})$
"""
- # TODO: Implement for Task 1.1.
- raise NotImplementedError("Need to implement for Task 1.1")
+ vals_pos = list(vals)
+ vals_neg = list(vals)
+
+ vals_pos[arg] = vals[arg] + epsilon
+ vals_neg[arg] = vals[arg] - epsilon
+
+ f_pos = f(*vals_pos)
+ f_neg = f(*vals_neg)
+
+ return (f_pos - f_neg) / (2.0 * epsilon)
variable_count = 1
@@ -61,8 +69,22 @@ def topological_sort(variable: Variable) -> Iterable[Variable]:
Returns:
Non-constant Variables in topological order starting from the right.
"""
- # TODO: Implement for Task 1.4.
- raise NotImplementedError("Need to implement for Task 1.4")
+ visited = set()
+ post: List[Variable] = []
+
+ def dfs(v: Variable) -> None:
+ uid = v.unique_id
+ if uid in visited:
+ return
+ visited.add(uid)
+ if v.is_constant():
+ return
+ for p in v.parents:
+ dfs(p)
+ post.append(v)
+
+ dfs(variable)
+ return list(reversed(post))
def backpropagate(variable: Variable, deriv: Any) -> None:
@@ -76,8 +98,20 @@ def backpropagate(variable: Variable, deriv: Any) -> None:
No return. Should write to its results to the derivative values of each leaf through `accumulate_derivative`.
"""
- # TODO: Implement for Task 1.4.
- raise NotImplementedError("Need to implement for Task 1.4")
+ topo = list(topological_sort(variable))
+
+ grads: dict[int, float] = {}
+ grads[variable.unique_id] = float(deriv)
+
+ for v in topo:
+ g_out = float(grads.get(v.unique_id, 0.0))
+ if v.is_leaf():
+ v.accumulate_derivative(g_out)
+ continue
+
+ for parent, g_local in v.chain_rule(g_out):
+ pid = parent.unique_id
+ grads[pid] = grads.get(pid, 0.0) + float(g_local)
@dataclass
diff --git a/minitorch/module.py b/minitorch/module.py
index 11fc1f39..dcdcffab 100644
--- a/minitorch/module.py
+++ b/minitorch/module.py
@@ -1,14 +1,14 @@
from __future__ import annotations
-from typing import Any, Dict, Optional, Sequence, Tuple
+from typing import Any, Dict, Optional, Sequence, Tuple, List
class Module:
- """
- Modules form a tree that store parameters and other
+ """Modules form a tree that store parameters and other
submodules. They make up the basis of neural network stacks.
- Attributes:
+ Attributes
+ ----------
_modules : Storage of the child modules
_parameters : Storage of the module's parameters
training : Whether the module is in training mode or evaluation mode
@@ -25,42 +25,57 @@ def __init__(self) -> None:
self.training = True
def modules(self) -> Sequence[Module]:
- "Return the direct child modules of this module."
+ """Return the direct child modules of this module."""
m: Dict[str, Module] = self.__dict__["_modules"]
return list(m.values())
def train(self) -> None:
- "Set the mode of this module and all descendent modules to `train`."
- raise NotImplementedError("Need to include this file from past assignment.")
+ """Set the mode of this module and all descendent modules to `train`."""
+ self.training = True
+ for child in self._modules.values():
+ child.train()
def eval(self) -> None:
- "Set the mode of this module and all descendent modules to `eval`."
- raise NotImplementedError("Need to include this file from past assignment.")
+ """Set the mode of this module and all descendent modules to `eval`."""
+ self.training = False
+ for child in self._modules.values():
+ child.eval()
def named_parameters(self) -> Sequence[Tuple[str, Parameter]]:
- """
- Collect all the parameters of this module and its descendents.
+ """Collect all the parameters of this module and its descendents.
-
- Returns:
+ Returns
+ -------
The name and `Parameter` of each ancestor parameter.
+
"""
- raise NotImplementedError("Need to include this file from past assignment.")
+ out: List[Tuple[str, Parameter]] = []
+
+ for name, p in self._parameters.items():
+ out.append((name, p))
+
+ for child_name, child_mod in self._modules.items():
+ for sub_name, p in child_mod.named_parameters():
+ out.append((f"{child_name}.{sub_name}", p))
+
+ return out
def parameters(self) -> Sequence[Parameter]:
- "Enumerate over all the parameters of this module and its descendents."
- raise NotImplementedError("Need to include this file from past assignment.")
+ """Enumerate over all the parameters of this module and its descendents."""
+ return [p for _, p in self.named_parameters()]
def add_parameter(self, k: str, v: Any) -> Parameter:
- """
- Manually add a parameter. Useful helper for scalar parameters.
+ """Manually add a parameter. Useful helper for scalar parameters.
Args:
+ ----
k: Local name of the parameter.
v: Value for the parameter.
Returns:
+ -------
Newly created parameter.
+
"""
val = Parameter(v, k)
self.__dict__["_parameters"][k] = val
@@ -114,8 +129,7 @@ def _addindent(s_: str, numSpaces: int) -> str:
class Parameter:
- """
- A Parameter is a special container stored in a `Module`.
+ """A Parameter is a special container stored in a `Module`.
It is designed to hold a `Variable`, but we allow it to hold
any value for testing.
@@ -130,7 +144,7 @@ def __init__(self, x: Any, name: Optional[str] = None) -> None:
self.value.name = self.name
def update(self, x: Any) -> None:
- "Update the parameter value."
+ """Update the parameter value."""
self.value = x
if hasattr(x, "requires_grad_"):
self.value.requires_grad_(True)
diff --git a/minitorch/operators.py b/minitorch/operators.py
index 895ae82d..efba54d9 100644
--- a/minitorch/operators.py
+++ b/minitorch/operators.py
@@ -1,185 +1,164 @@
-"""
-Collection of the core mathematical operators used throughout the code base.
-"""
+"""Collection of the core mathematical operators used throughout the code base."""
import math
-from typing import Callable, Iterable
# ## Task 0.1
+from typing import Callable, Iterable, Iterator, List, TypeVar, Sequence
+
#
# Implementation of a prelude of elementary functions.
+# Mathematical functions:
+# - mul
+# - id
+# - add
+# - neg
+# - lt
+# - eq
+# - max
+# - is_close
+# - sigmoid
+# - relu
+# - log
+# - exp
+# - log_back
+# - inv
+# - inv_back
+# - relu_back
+#
+# For sigmoid calculate as:
+# $f(x) = \frac{1.0}{(1.0 + e^{-x})}$ if x >=0 else $\frac{e^x}{(1.0 + e^{x})}$
+# For is_close:
+# $f(x) = |x - y| < 1e-2$
+
def mul(x: float, y: float) -> float:
- "$f(x, y) = x * y$"
- raise NotImplementedError("Need to include this file from past assignment.")
+ return x * y
def id(x: float) -> float:
- "$f(x) = x$"
- raise NotImplementedError("Need to include this file from past assignment.")
+ return x
def add(x: float, y: float) -> float:
- "$f(x, y) = x + y$"
- raise NotImplementedError("Need to include this file from past assignment.")
+ return x + y
def neg(x: float) -> float:
- "$f(x) = -x$"
- raise NotImplementedError("Need to include this file from past assignment.")
+ return -x
def lt(x: float, y: float) -> float:
- "$f(x) =$ 1.0 if x is less than y else 0.0"
- raise NotImplementedError("Need to include this file from past assignment.")
+ return 1.0 if x < y else 0.0
def eq(x: float, y: float) -> float:
- "$f(x) =$ 1.0 if x is equal to y else 0.0"
- raise NotImplementedError("Need to include this file from past assignment.")
+ return 1.0 if x == y else 0.0
def max(x: float, y: float) -> float:
- "$f(x) =$ x if x is greater than y else y"
- raise NotImplementedError("Need to include this file from past assignment.")
+ return x if x > y else y
-def is_close(x: float, y: float) -> float:
- "$f(x) = |x - y| < 1e-2$"
- raise NotImplementedError("Need to include this file from past assignment.")
+def is_close(x: float, y: float) -> bool:
+ return abs(x - y) < 1e-2
def sigmoid(x: float) -> float:
- r"""
- $f(x) = \frac{1.0}{(1.0 + e^{-x})}$
-
- (See https://en.wikipedia.org/wiki/Sigmoid_function )
-
- Calculate as
-
- $f(x) = \frac{1.0}{(1.0 + e^{-x})}$ if x >=0 else $\frac{e^x}{(1.0 + e^{x})}$
-
- for stability.
- """
- raise NotImplementedError("Need to include this file from past assignment.")
+ if x >= 0.0:
+ z = math.exp(-x)
+ return 1.0 / (1.0 + z)
+ else:
+ z = math.exp(x)
+ return z / (1.0 + z)
def relu(x: float) -> float:
- """
- $f(x) =$ x if x is greater than 0, else 0
-
- (See https://en.wikipedia.org/wiki/Rectifier_(neural_networks) .)
- """
- raise NotImplementedError("Need to include this file from past assignment.")
-
-
-EPS = 1e-6
+ return x if x > 0.0 else 0.0
def log(x: float) -> float:
- "$f(x) = log(x)$"
- return math.log(x + EPS)
+ return math.log(x)
def exp(x: float) -> float:
- "$f(x) = e^{x}$"
return math.exp(x)
-def log_back(x: float, d: float) -> float:
- r"If $f = log$ as above, compute $d \times f'(x)$"
- raise NotImplementedError("Need to include this file from past assignment.")
+def inv(x: float) -> float:
+ return 1.0 / x
-def inv(x: float) -> float:
- "$f(x) = 1/x$"
- raise NotImplementedError("Need to include this file from past assignment.")
+def log_back(a: float, b: float) -> float:
+ return b / a
-def inv_back(x: float, d: float) -> float:
- r"If $f(x) = 1/x$ compute $d \times f'(x)$"
- raise NotImplementedError("Need to include this file from past assignment.")
+def inv_back(a: float, b: float) -> float:
+ return -b / (a * a)
-def relu_back(x: float, d: float) -> float:
- r"If $f = relu$ compute $d \times f'(x)$"
- raise NotImplementedError("Need to include this file from past assignment.")
+def relu_back(a: float, b: float) -> float:
+ return b if a > 0.0 else 0.0
# ## Task 0.3
# Small practice library of elementary higher-order functions.
+# Implement the following core functions
+# - map
+# - zipWith
+# - reduce
+#
+# Use these to implement
+# - negList : negate a list
+# - addLists : add two lists together
+# - sum: sum lists
+# - prod: take the product of lists
-def map(fn: Callable[[float], float]) -> Callable[[Iterable[float]], Iterable[float]]:
- """
- Higher-order map.
-
- See https://en.wikipedia.org/wiki/Map_(higher-order_function)
-
- Args:
- fn: Function from one value to one value.
-
- Returns:
- A function that takes a list, applies `fn` to each element, and returns a
- new list
- """
- raise NotImplementedError("Need to include this file from past assignment.")
-
-
-def negList(ls: Iterable[float]) -> Iterable[float]:
- "Use `map` and `neg` to negate each element in `ls`"
- raise NotImplementedError("Need to include this file from past assignment.")
-
-
-def zipWith(
- fn: Callable[[float, float], float]
-) -> Callable[[Iterable[float], Iterable[float]], Iterable[float]]:
- """
- Higher-order zipwith (or map2).
- See https://en.wikipedia.org/wiki/Map_(higher-order_function)
+T = TypeVar("T")
+U = TypeVar("U")
- Args:
- fn: combine two values
- Returns:
- Function that takes two equally sized lists `ls1` and `ls2`, produce a new list by
- applying fn(x, y) on each pair of elements.
+def map(fn: Callable[[T], U], it: Iterable[T]) -> List[U]:
+ out: List[U] = []
+ for v in it:
+ out.append(fn(v))
+ return out
- """
- raise NotImplementedError("Need to include this file from past assignment.")
+def zipWith(fn: Callable[[T, U], T], a: Iterable[T], b: Iterable[U]) -> List[T]:
+ out: List[T] = []
+ ia = iter(a)
+ ib = iter(b)
+ while True:
+ try:
+ va = next(ia)
+ vb = next(ib)
+ except StopIteration:
+ break
+ out.append(fn(va, vb))
+ return out
-def addLists(ls1: Iterable[float], ls2: Iterable[float]) -> Iterable[float]:
- "Add the elements of `ls1` and `ls2` using `zipWith` and `add`"
- raise NotImplementedError("Need to include this file from past assignment.")
+def reduce(fn: Callable[[T, T], T], it: Iterable[T], start: T) -> T:
+ acc: T = start
+ for v in it:
+ acc = fn(acc, v)
+ return acc
-def reduce(
- fn: Callable[[float, float], float], start: float
-) -> Callable[[Iterable[float]], float]:
- r"""
- Higher-order reduce.
+def negList(ls: Iterable[float]) -> List[float]:
+ return map(neg, ls)
- Args:
- fn: combine two values
- start: start value $x_0$
- Returns:
- Function that takes a list `ls` of elements
- $x_1 \ldots x_n$ and computes the reduction :math:`fn(x_3, fn(x_2,
- fn(x_1, x_0)))`
- """
- raise NotImplementedError("Need to include this file from past assignment.")
+def addLists(a: Iterable[float], b: Iterable[float]) -> List[float]:
+ return zipWith(add, a, b)
def sum(ls: Iterable[float]) -> float:
- "Sum up a list using `reduce` and `add`."
- raise NotImplementedError("Need to include this file from past assignment.")
+ return reduce(add, ls, 0.0)
def prod(ls: Iterable[float]) -> float:
- "Product of a list using `reduce` and `mul`."
- raise NotImplementedError("Need to include this file from past assignment.")
+ return reduce(mul, ls, 1.0)
diff --git a/minitorch/scalar.py b/minitorch/scalar.py
index f5abbe9e..6e4b1ec6 100644
--- a/minitorch/scalar.py
+++ b/minitorch/scalar.py
@@ -92,31 +92,25 @@ def __rtruediv__(self, b: ScalarLike) -> Scalar:
return Mul.apply(b, Inv.apply(self))
def __add__(self, b: ScalarLike) -> Scalar:
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
+ return Add.apply(self, b)
def __bool__(self) -> bool:
return bool(self.data)
def __lt__(self, b: ScalarLike) -> Scalar:
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
+ return LT.apply(self, b)
def __gt__(self, b: ScalarLike) -> Scalar:
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
+ return LT.apply(b, self)
def __eq__(self, b: ScalarLike) -> Scalar: # type: ignore[override]
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
+ return EQ.apply(self, b)
def __sub__(self, b: ScalarLike) -> Scalar:
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
+ return Add.apply(self, Neg.apply(b))
def __neg__(self) -> Scalar:
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
+ return Neg.apply(self)
def __radd__(self, b: ScalarLike) -> Scalar:
return self + b
@@ -125,20 +119,16 @@ def __rmul__(self, b: ScalarLike) -> Scalar:
return self * b
def log(self) -> Scalar:
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
+ return Log.apply(self)
def exp(self) -> Scalar:
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
+ return Exp.apply(self)
def sigmoid(self) -> Scalar:
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
+ return Sigmoid.apply(self)
def relu(self) -> Scalar:
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
+ return ReLU.apply(self)
# Variable elements for backprop
@@ -173,8 +163,13 @@ def chain_rule(self, d_output: Any) -> Iterable[Tuple[Variable, Any]]:
assert h.last_fn is not None
assert h.ctx is not None
- # TODO: Implement for Task 1.3.
- raise NotImplementedError("Need to implement for Task 1.3")
+ local_grads = h.last_fn._backward(h.ctx, d_output)
+
+ out: list[tuple[Variable, Any]] = []
+ for inp, g in zip(h.inputs, local_grads):
+ if not inp.is_constant():
+ out.append((inp, g))
+ return out
def backward(self, d_output: Optional[float] = None) -> None:
"""
diff --git a/minitorch/scalar_functions.py b/minitorch/scalar_functions.py
index d8d2307b..e9e1f5e1 100644
--- a/minitorch/scalar_functions.py
+++ b/minitorch/scalar_functions.py
@@ -103,13 +103,13 @@ class Mul(ScalarFunction):
@staticmethod
def forward(ctx: Context, a: float, b: float) -> float:
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
-
+ ctx.save_for_backward(a, b)
+ return float(operators.mul(a, b))
+
@staticmethod
def backward(ctx: Context, d_output: float) -> Tuple[float, float]:
- # TODO: Implement for Task 1.4.
- raise NotImplementedError("Need to implement for Task 1.4")
+ a, b = ctx.saved_values
+ return float(d_output * b), float(d_output * a)
class Inv(ScalarFunction):
@@ -117,13 +117,13 @@ class Inv(ScalarFunction):
@staticmethod
def forward(ctx: Context, a: float) -> float:
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
+ ctx.save_for_backward(a)
+ return float(operators.inv(a))
@staticmethod
def backward(ctx: Context, d_output: float) -> float:
- # TODO: Implement for Task 1.4.
- raise NotImplementedError("Need to implement for Task 1.4")
+ (a,) = ctx.saved_values
+ return float(operators.inv_back(a, d_output))
class Neg(ScalarFunction):
@@ -131,13 +131,11 @@ class Neg(ScalarFunction):
@staticmethod
def forward(ctx: Context, a: float) -> float:
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
+ return float(operators.neg(a))
@staticmethod
def backward(ctx: Context, d_output: float) -> float:
- # TODO: Implement for Task 1.4.
- raise NotImplementedError("Need to implement for Task 1.4")
+ return float(-d_output)
class Sigmoid(ScalarFunction):
@@ -145,13 +143,14 @@ class Sigmoid(ScalarFunction):
@staticmethod
def forward(ctx: Context, a: float) -> float:
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
+ s = operators.sigmoid(a)
+ ctx.save_for_backward(s)
+ return float(s)
@staticmethod
def backward(ctx: Context, d_output: float) -> float:
- # TODO: Implement for Task 1.4.
- raise NotImplementedError("Need to implement for Task 1.4")
+ (s,) = ctx.saved_values
+ return float(d_output * s * (1.0 - s))
class ReLU(ScalarFunction):
@@ -159,13 +158,13 @@ class ReLU(ScalarFunction):
@staticmethod
def forward(ctx: Context, a: float) -> float:
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
+ ctx.save_for_backward(a)
+ return float(operators.relu(a))
@staticmethod
def backward(ctx: Context, d_output: float) -> float:
- # TODO: Implement for Task 1.4.
- raise NotImplementedError("Need to implement for Task 1.4")
+ (a,) = ctx.saved_values
+ return float(operators.relu_back(a, d_output))
class Exp(ScalarFunction):
@@ -173,13 +172,14 @@ class Exp(ScalarFunction):
@staticmethod
def forward(ctx: Context, a: float) -> float:
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
+ e = operators.exp(a)
+ ctx.save_for_backward(e)
+ return float(e)
@staticmethod
def backward(ctx: Context, d_output: float) -> float:
- # TODO: Implement for Task 1.4.
- raise NotImplementedError("Need to implement for Task 1.4")
+ (e,) = ctx.saved_values
+ return float(d_output * e)
class LT(ScalarFunction):
@@ -187,13 +187,11 @@ class LT(ScalarFunction):
@staticmethod
def forward(ctx: Context, a: float, b: float) -> float:
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
+ return float(operators.lt(a, b))
@staticmethod
def backward(ctx: Context, d_output: float) -> Tuple[float, float]:
- # TODO: Implement for Task 1.4.
- raise NotImplementedError("Need to implement for Task 1.4")
+ return 0.0, 0.0
class EQ(ScalarFunction):
@@ -201,10 +199,8 @@ class EQ(ScalarFunction):
@staticmethod
def forward(ctx: Context, a: float, b: float) -> float:
- # TODO: Implement for Task 1.2.
- raise NotImplementedError("Need to implement for Task 1.2")
+ return float(operators.eq(a, b))
@staticmethod
def backward(ctx: Context, d_output: float) -> Tuple[float, float]:
- # TODO: Implement for Task 1.4.
- raise NotImplementedError("Need to implement for Task 1.4")
+ return 0.0, 0.0
diff --git a/project/run_scalar.py b/project/run_scalar.py
index 7ce5207b..2afef78b 100644
--- a/project/run_scalar.py
+++ b/project/run_scalar.py
@@ -10,8 +10,9 @@
class Network(minitorch.Module):
def __init__(self, hidden_layers):
super().__init__()
- # TODO: Implement for Task 1.5.
- raise NotImplementedError("Need to implement for Task 1.5")
+ self.layer1 = Linear(in_size=2, out_size=hidden_layers)
+ self.layer2 = Linear(in_size=hidden_layers, out_size=hidden_layers)
+ self.layer3 = Linear(in_size=hidden_layers, out_size=1)
def forward(self, x):
middle = [h.relu() for h in self.layer1.forward(x)]
@@ -40,9 +41,15 @@ def __init__(self, in_size, out_size):
)
def forward(self, inputs):
- # TODO: Implement for Task 1.5.
- raise NotImplementedError("Need to implement for Task 1.5")
-
+ xs = list(inputs)
+ out = []
+ out_size = len(self.bias)
+ for j in range(out_size):
+ s = self.bias[j].value
+ for i, x in enumerate(xs):
+ s = s + x * self.weights[i][j].value
+ out.append(s)
+ return out
def default_log_fn(epoch, total_loss, correct, losses):
print("Epoch ", epoch, " loss ", total_loss, "correct", correct)
@@ -101,7 +108,7 @@ def train(self, data, learning_rate, max_epochs=500, log_fn=default_log_fn):
if __name__ == "__main__":
PTS = 50
- HIDDEN = 2
+ HIDDEN = 6
RATE = 0.5
- data = minitorch.datasets["Simple"](PTS)
- ScalarTrain(HIDDEN).train(data, RATE)
+ data = minitorch.datasets["Split"](PTS)
+ ScalarTrain(HIDDEN).train(data, RATE, max_epochs=800)
\ No newline at end of file
From 76657090199d5a6d6dee646703f4ff195d723675 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=D0=92=D0=B0=D0=B4=D0=B8=D0=BC=20=D0=9F=D0=B0=D0=BB=D1=8B?=
=?UTF-8?q?=D1=81=D0=B0=D0=B5=D0=B2?=
Date: Sun, 19 Oct 2025 23:45:10 +0300
Subject: [PATCH 6/8] workflows
---
.github/workflows/minitorch.yml | 40 +++++++++++++++++++++++++++++++++
1 file changed, 40 insertions(+)
create mode 100644 .github/workflows/minitorch.yml
diff --git a/.github/workflows/minitorch.yml b/.github/workflows/minitorch.yml
new file mode 100644
index 00000000..923052da
--- /dev/null
+++ b/.github/workflows/minitorch.yml
@@ -0,0 +1,40 @@
+name: CI (Module 1)
+
+on:
+ push:
+ pull_request:
+
+jobs:
+ tests:
+ runs-on: ubuntu-latest
+ steps:
+ - name: Checkout
+ uses: actions/checkout@v4
+
+ - name: Set up Python
+ uses: actions/setup-python@v5
+ with:
+ python-version: "3.11"
+ cache: "pip"
+
+ - name: Install dependencies
+ run: |
+ sudo apt-get update
+ sudo apt-get install -y graphviz
+ python -m pip install --upgrade pip
+ if [ -f requirements.txt ]; then pip install -r requirements.txt; fi
+ pip install -e .
+ pip install pytest hypothesis flake8 pep8-naming
+
+ - name: Lint with flake8
+ run: |
+ # если есть .flake8 в корне — flake8 прочитает правила из него
+ flake8 minitorch/ tests/ project/
+
+ - name: Run tests (Module 1 only)
+ run: |
+ echo "Module 1"
+ pytest -q -m task1_1
+ pytest -q -m task1_2
+ pytest -q -m task1_3
+ pytest -q -m task1_4
From 3d29208d9581364f0a07f2203068cfd8be21ae41 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=D0=92=D0=B0=D0=B4=D0=B8=D0=BC=20=D0=9F=D0=B0=D0=BB=D1=8B?=
=?UTF-8?q?=D1=81=D0=B0=D0=B5=D0=B2?=
Date: Sun, 19 Oct 2025 23:50:42 +0300
Subject: [PATCH 7/8] workflows
---
.github/workflows/minitorch.yml | 29 +++++++++++++++++------------
1 file changed, 17 insertions(+), 12 deletions(-)
diff --git a/.github/workflows/minitorch.yml b/.github/workflows/minitorch.yml
index 923052da..776051ae 100644
--- a/.github/workflows/minitorch.yml
+++ b/.github/workflows/minitorch.yml
@@ -5,36 +5,41 @@ on:
pull_request:
jobs:
- tests:
+ build:
runs-on: ubuntu-latest
+ strategy:
+ matrix:
+ python-version: ["3.10"]
+
steps:
- name: Checkout
uses: actions/checkout@v4
- - name: Set up Python
+ - name: Set up Python ${{ matrix.python-version }}
uses: actions/setup-python@v5
with:
- python-version: "3.11"
- cache: "pip"
+ python-version: ${{ matrix.python-version }}
+ cache: pip
- name: Install dependencies
run: |
sudo apt-get update
sudo apt-get install -y graphviz
python -m pip install --upgrade pip
+ pip install flake8 pytest pep8-naming
if [ -f requirements.txt ]; then pip install -r requirements.txt; fi
+ # на всякий случай — установить сам пакет
pip install -e .
- pip install pytest hypothesis flake8 pep8-naming
- name: Lint with flake8
run: |
- # если есть .flake8 в корне — flake8 прочитает правила из него
- flake8 minitorch/ tests/ project/
+ # stop the build if there are Python syntax errors or undefined names
+ flake8 --ignore "N801,E203,E266,E501,W503,W504,F812,F401,F841,E741,N803,N802,N806,E128,E302" minitorch/ tests/ project/
- - name: Run tests (Module 1 only)
+ - name: Test with pytest (Module 1)
run: |
echo "Module 1"
- pytest -q -m task1_1
- pytest -q -m task1_2
- pytest -q -m task1_3
- pytest -q -m task1_4
+ pytest tests -q -x -m task1_1
+ pytest tests -q -x -m task1_2
+ pytest tests -q -x -m task1_3
+ pytest tests -q -x -m task1_4
From bd4c2030e12168b45ddaae90309cde5855a2e7cb Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=D0=92=D0=B0=D0=B4=D0=B8=D0=BC=20=D0=9F=D0=B0=D0=BB=D1=8B?=
=?UTF-8?q?=D1=81=D0=B0=D0=B5=D0=B2?=
Date: Sun, 19 Oct 2025 23:54:56 +0300
Subject: [PATCH 8/8] fix linker
---
minitorch/scalar_functions.py | 2 +-
project/run_scalar.py | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/minitorch/scalar_functions.py b/minitorch/scalar_functions.py
index e9e1f5e1..92ee9ea2 100644
--- a/minitorch/scalar_functions.py
+++ b/minitorch/scalar_functions.py
@@ -105,7 +105,7 @@ class Mul(ScalarFunction):
def forward(ctx: Context, a: float, b: float) -> float:
ctx.save_for_backward(a, b)
return float(operators.mul(a, b))
-
+
@staticmethod
def backward(ctx: Context, d_output: float) -> Tuple[float, float]:
a, b = ctx.saved_values
diff --git a/project/run_scalar.py b/project/run_scalar.py
index 2afef78b..26703cb0 100644
--- a/project/run_scalar.py
+++ b/project/run_scalar.py
@@ -111,4 +111,4 @@ def train(self, data, learning_rate, max_epochs=500, log_fn=default_log_fn):
HIDDEN = 6
RATE = 0.5
data = minitorch.datasets["Split"](PTS)
- ScalarTrain(HIDDEN).train(data, RATE, max_epochs=800)
\ No newline at end of file
+ ScalarTrain(HIDDEN).train(data, RATE, max_epochs=800)