Hugging Face's logo

CompassioninMachineLearning
/
olmo-compassion-part1

Safetensors
olmo3
Model card Files
xet
 Community
olmo-compassion-part1
File size: 2,503 Bytes
06aa8e4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
{
 "run_name": "olmo-compassion-part1",
 "git_sha": "f56d0c01687b3b599f164c8b826c3df524c66dbf",
 "argv": [
  "pilot/train_unsloth.py",
  "--data",
  "/workspace/anchor/olmo-compassion/train_olmo-compassion_replay50.jsonl",
  "/workspace/data/xlam5000_max1call.jsonl",
  "--base-model",
  "CompassioninMachineLearning/Olmo7b-compassion-cleaned-10k-CPT-best-step750",
  "--template-source",
  "allenai/Olmo-3-7B-Instruct",
  "--no-4bit",
  "--lora-r",
  "64",
  "--lora-alpha",
  "128",
  "--seed",
  "7",
  "--epochs",
  "1",
  "--eval-steps",
  "50",
  "--early-stopping-patience",
  "2",
  "--snapshot-frac",
  "0",
  "--replay-chat",
  "--replay-token-frac",
  "0.25",
  "--run-name",
  "olmo-compassion-part1",
  "--push-to-hub-merged",
  "CompassioninMachineLearning/olmo-compassion-part1-build",
  "--output-dir",
  "/workspace/anchor/olmo-compassion/train"
 ],
 "base_model": "CompassioninMachineLearning/Olmo7b-compassion-cleaned-10k-CPT-best-step750",
 "template_source": "allenai/Olmo-3-7B-Instruct",
 "modules_to_save": [
  "embed_tokens",
  "lm_head"
 ],
 "data_files": [
  "/workspace/anchor/olmo-compassion/train_olmo-compassion_replay50.jsonl",
  "/workspace/data/xlam5000_max1call.jsonl"
 ],
 "data_sha256": {
  "/workspace/anchor/olmo-compassion/train_olmo-compassion_replay50.jsonl": "b43d87068505b5bbdb7193c1633f241ffa36e3aeb6b940fea872d3a7fda69226",
  "/workspace/data/xlam5000_max1call.jsonl": "a8aae6cf5c36dcb61b311fed62a8a5b61674e0aa6a3929ab7ddcbf4a7389b5cd"
 },
 "rows_raw": 23500,
 "rows_kept": 15331,
 "by_source": {
  "recovery": 3750,
  "xlam": 5000,
  "replay:compassion_12185_cleaned": 1257,
  "apigen": 4824,
  "efficiency": 500
 },
 "train_examples": 14947,
 "eval_examples": 384,
 "hyperparams": {
  "lora_r": 64,
  "lora_alpha": 128,
  "lora_dropout": 0.05,
  "lr_adapters": 0.0002,
  "lr_embeddings": 2e-05,
  "epochs": 1.0,
  "max_steps": -1,
  "max_seq_len": 8192,
  "effective_batch": 16,
  "per_device_batch": 2,
  "grad_accum": 8,
  "warmup_ratio": 0.05,
  "schedule": "cosine",
  "seed": 7,
  "snapshot_frac": 0.0
 },
 "versions": {
  "torch": "2.11.0+cu128",
  "transformers": "4.57.6",
  "unsloth": "2026.9.2",
  "trl": "0.22.2",
  "datasets": "5.0.1",
  "peft": "0.20.0",
  "accelerate": "1.14.0",
  "bitsandbytes": "0.50.2"
 },
 "gpu": "NVIDIA H100 80GB HBM3",
 "started_at": "2026-10-13T00:08:22Z",
 "exported_checkpoint": "checkpoint-200",
 "exported_step": 200,
 "export_rule": "common step: the earliest step at which every Olmo anchor passes GATE_OLMO"
}