config: set Stage 1 micro 32 accum 4 for peak Tensor Core speed 8866f9e verified ViuAI commited on about 1 hour ago
config: set micro 16 accum 8 with grad_ckpt true for safe high-throughput pretraining 97527fa verified ViuAI commited on about 1 hour ago
fix: restore grad_ckpt and safe micro 8 accum 8 for zero-OOM guarantee on T4 96be75f verified ViuAI commited on about 1 hour ago
perf: disable grad_ckpt on T4 and enlarge loss chunk_size to 2048 for 10k+ tok/s 7d4c3f5 verified ViuAI commited on about 1 hour ago
config: update Stage 1 batch_size to 16 for micro=16 accum=8 per GPU 20ebd23 verified ViuAI commited on about 1 hour ago
refactor: consolidate pure 32L Dense architecture, purge MoE files, audit hardening 3a5d3e5 verified ViuAI commited on about 2 hours ago
Update docs/KAGGLE_RUN.md: log_every: 10 and real-time step streaming f185a5b verified ViuAI commited on about 2 hours ago
Update model/configs/train_kaggle_t4.yaml: log_every: 10 and real-time step streaming a68274b verified ViuAI commited on about 2 hours ago
Remove dead config model/configs/model_config.yaml f3b7a40 verified ViuAI commited on about 3 hours ago
Remove dead config model/configs/viu_1_5b_moe_config.yaml c4713b4 verified ViuAI commited on about 3 hours ago
Remove dead config model/configs/viu1_moe_config.yaml ff9965d verified ViuAI commited on about 3 hours ago
Archive viu_1_5b_moe_config.yaml to deprecated folder 96ebcea verified ViuAI commited on about 3 hours ago
Archive viu1_moe_config.yaml to deprecated folder e910481 verified ViuAI commited on about 3 hours ago
Fix model/scripts/train.py: batch math comments and find_unused_parameters=False ef65129 verified ViuAI commited on about 3 hours ago
Fix model/configs/train_config.yaml: batch math comments and find_unused_parameters=False d7975ca verified ViuAI commited on about 3 hours ago
Audit fix: Update tokenizer/configs/tokenizer_config.json for ViuMini-Dense-360M 8ab3701 verified ViuAI commited on about 4 hours ago
Audit fix: Update tokenizer/configs/tokenizer_config.yaml for ViuMini-Dense-360M 91f527b verified ViuAI commited on about 4 hours ago
Audit fix: Update model/scripts/train.py for ViuMini-Dense-360M e52f2b4 verified ViuAI commited on about 4 hours ago
Audit fix: Update model/configs/viu_1_5b_moe_config.yaml for ViuMini-Dense-360M 5d2e7a3 verified ViuAI commited on about 4 hours ago
Audit fix: Update model/configs/viu1_moe_config.yaml for ViuMini-Dense-360M a072aa0 verified ViuAI commited on about 4 hours ago
Audit fix: Update model/configs/model_config.yaml for ViuMini-Dense-360M d32d6e4 verified ViuAI commited on about 4 hours ago
Audit fix: Update model/configs/train_dense_baseline.yaml for ViuMini-Dense-360M 41869c6 verified ViuAI commited on about 4 hours ago
Audit fix: Update model/configs/train_h100_h200.yaml for ViuMini-Dense-360M 232a8e4 verified ViuAI commited on about 4 hours ago
Audit fix: Update model/configs/train_rtx5090.yaml for ViuMini-Dense-360M 70b87d7 verified ViuAI commited on about 4 hours ago
Audit fix: Update model/configs/train_rtx4090.yaml for ViuMini-Dense-360M d33c293 verified ViuAI commited on about 4 hours ago
Audit fix: Update model/configs/train_rtx3090.yaml for ViuMini-Dense-360M da2ff29 verified ViuAI commited on about 4 hours ago
Audit fix: Update model/configs/config.json for ViuMini-Dense-360M 6e8fe69 verified ViuAI commited on about 4 hours ago
Audit fix: Update model/configs/viu1_dense_config.yaml for ViuMini-Dense-360M d8ba5af verified ViuAI commited on about 4 hours ago
Audit fix: Update model/NOTES.md for ViuMini-Dense-360M 42788da verified ViuAI commited on about 4 hours ago
Audit fix: Update docs/PROGRESS.md for ViuMini-Dense-360M 268f804 verified ViuAI commited on about 4 hours ago
Audit fix: Update docs/KAGGLE_RUN.md for ViuMini-Dense-360M f859acd verified ViuAI commited on about 4 hours ago
Audit fix: Update requirements.txt for ViuMini-Dense-360M d248f4a verified ViuAI commited on about 4 hours ago
Audit fix: Update NOTES.md for ViuMini-Dense-360M bd55c48 verified ViuAI commited on about 4 hours ago
Audit fix: Update README.md for ViuMini-Dense-360M 1d7d841 verified ViuAI commited on about 4 hours ago
Update model/scripts/push_to_hf.py for ViuMini-Dense-360M d9702de verified ViuAI commited on about 5 hours ago
Update model/scripts/train.py for ViuMini-Dense-360M dedfdc7 verified ViuAI commited on about 5 hours ago
Update model/configs/train_rtx5090.yaml for ViuMini-Dense-360M a2f494b verified ViuAI commited on about 5 hours ago
Update model/configs/train_rtx4090.yaml for ViuMini-Dense-360M c8a4a4d verified ViuAI commited on about 5 hours ago
Update model/configs/train_rtx3090.yaml for ViuMini-Dense-360M ee5595d verified ViuAI commited on about 5 hours ago
Update model/configs/train_dense_baseline.yaml for ViuMini-Dense-360M 08fb77d verified ViuAI commited on about 5 hours ago
Update model/configs/train_config.yaml for ViuMini-Dense-360M 6a651bd verified ViuAI commited on about 5 hours ago
Update model/configs/train_kaggle_t4.yaml for ViuMini-Dense-360M c01244c verified ViuAI commited on about 5 hours ago
Update docs/HF_PUSH_GUIDE.md for ViuMini-Dense-360M 1a95c50 verified ViuAI commited on about 5 hours ago
Update docs/KAGGLE_RUN.md for ViuMini-Dense-360M c49f48e verified ViuAI commited on about 5 hours ago
Update NOTES.md for ViuMini-Dense-360M 32L architecture 0027325 verified ViuAI commited on about 5 hours ago
Update README.md for ViuMini-Dense-360M 32L architecture 333904e verified ViuAI commited on about 5 hours ago
Set Option A: 32 layers, dim 960 (366.6M params, SmolLM-style) 85ae4bc verified ViuAI commited on about 5 hours ago