Run smoke test to validate all fixes before full training #3

New issue

Open

opened 2026-05-01 14:01:26 +02:00 by sleepy · 0 comments

sleepy commented 2026-05-01 14:01:26 +02:00 (Migrated from localhost:18431)

Copy link

Goal

Verify that the fixed modules produce reasonable loss values and gradients flow correctly.

Test (fixes/smoke_test.py)

• Forward pass: loss should be < 10.4 (log(32K vocab))
• Backward pass: gradients should exist and be non-zero
• RoPE: verify correct frequency interleaving
• Data alignment: input/labels match properly

Command

cd /home/sleepy/ternary && CUDA_VISIBLE_DEVICES=0 python fixes/train_fixed.py --smoke

Expected

• Initial loss ~10.4 (random), decreasing within first 50 steps
• No NaN/Inf in gradients
• RoPE test passes

Status

• Smoke test run and passed

## Goal Verify that the fixed modules produce reasonable loss values and gradients flow correctly. ## Test (fixes/smoke_test.py) - Forward pass: loss should be < 10.4 (log(32K vocab)) - Backward pass: gradients should exist and be non-zero - RoPE: verify correct frequency interleaving - Data alignment: input/labels match properly ## Command ```bash cd /home/sleepy/ternary && CUDA_VISIBLE_DEVICES=0 python fixes/train_fixed.py --smoke ``` ## Expected - Initial loss ~10.4 (random), decreasing within first 50 steps - No NaN/Inf in gradients - RoPE test passes ## Status - [ ] Smoke test run and passed

Sign in to join this conversation.

No Branch/Tag specified

Branches Tags

master

fix/1-integrate-fixes

fix/5-sweep-json-serialization

No results found.

Labels

Clear labels

No items

No labels

Milestone

Clear milestone

No items

No milestone

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

1 participant

Notifications

Subscribe

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

sleepy/ternary#3

No description provided.