-
7181acc580
fix: save train/test.jsonl to output_dir not parent
Christian Medina
2026-06-30 20:01:05 -04:00
-
d5ba471b98
fix: auto-split dataset if split files missing
Christian Medina
2026-06-30 19:54:50 -04:00
-
bbdbe91871
fix: fix dataset path calculation
Christian Medina
2026-06-30 19:01:39 -04:00
-
14a302dc5e
fix: make scripts executable
Christian Medina
2026-06-30 18:49:39 -04:00
-
97004261f5
fix: multiple bug fixes - remove unused imports, fix typo, fix dataset paths, add ai command fallback
Christian Medina
2026-06-30 18:48:46 -04:00
-
75aeeeb20a
fix: use torchrun for distributed training, load model on CPU
Christian Medina
2026-06-30 18:45:14 -04:00
-
5ad5518ca6
fix: add deepspeed to auto-install dependencies
Christian Medina
2026-06-30 18:30:23 -04:00
-
4f2f8ef03f
feat: use DeepSpeed ZeRO-3 for optimal 2-GPU training
Christian Medina
2026-06-30 18:26:48 -04:00
-
a04bf4a9cf
feat: add run_with_accelerate.sh helper script
Christian Medina
2026-06-30 18:26:08 -04:00
-
cd58ffd31e
fix: use FSDP for 2-GPU training, reduce batch size
Christian Medina
2026-06-30 18:23:34 -04:00
-
3d87f0b8b3
fix: force reinstall PyTorch to resolve version conflicts
Christian Medina
2026-06-30 17:58:52 -04:00
-
ef42e67183
fix: use CompressedTensors for Ornith, add torch import, expand LoRA targets
Christian Medina
2026-06-30 17:54:35 -04:00
-
7e8ddabf03
fix: remove dataset download, it's already in repo
Christian Medina
2026-06-30 17:11:29 -04:00
-
28d0a956fd
fix: update PyTorch to cu124 for RTX 5090, fix total_memory typo
Christian Medina
2026-06-30 16:58:11 -04:00
-
1786830d84
feat: unload AI model before training
Christian Medina
2026-06-30 16:00:17 -04:00
-
2074d3d993
feat: use Ornith-1.0-35B-FP8 with multi-GPU support
Christian Medina
2026-06-30 15:57:54 -04:00
-
215886d6a7
fix: use $HOME/loras by default
Christian Medina
2026-06-30 15:51:28 -04:00
-
11951a8276
feat: add deploy-agenx-lora.sh to repo
Christian Medina
2026-06-30 15:38:05 -04:00
-
39a493f261
fix: remove deploy-agenx-lora.sh from repo
Christian Medina
2026-06-30 15:35:42 -04:00
-
11653ed743
feat: add simple deploy-agenx-lora.sh script
Christian Medina
2026-06-30 15:31:13 -04:00
-
44939e374b
refactor: simplify deploy script, add train-on-this-server.sh
Christian Medina
2026-06-30 15:22:54 -04:00
-
0614166c7a
fix: use remote Gitea directly, remove local fallback
Christian Medina
2026-06-30 15:20:03 -04:00
-
6946ebac1a
feat: use local Gitea when available, fallback to remote
Christian Medina
2026-06-30 15:15:51 -04:00
-
8e51c39c6e
fix: remove large dataset from git tracking, add to .gitignore
Christian Medina
2026-06-30 15:06:13 -04:00
-
e1b80301d8
refactor: update deploy script to match cyron pattern
Christian Medina
2026-06-30 15:03:40 -04:00
-
3d2ae812a7
feat: add deploy-and-train.sh script for server deployment
Christian Medina
2026-06-30 15:02:13 -04:00
-
418a4cc76d
init: add LoRA training infrastructure and 20k dataset
Christian Medina
2026-06-30 14:49:44 -04:00