Hugging Face's logo

mlfoundations-dev
/
hp_ablations_qwen_scheduler_constant

Text Generation
Transformers
Safetensors
qwen2
llama-factory
full
Generated from Trainer
conversational
text-generation-inference
Inference Endpoints
Model card Files Community
sedrickkeh commited on
Commit
72cf443
•
1 Parent(s): 2edb3be

Training in progress, epoch 1

Browse files
Files changed (5) hide show
  1. model-00001-of-00004.safetensors +1 -1
  2. model-00002-of-00004.safetensors +1 -1
  3. model-00003-of-00004.safetensors +1 -1
  4. model-00004-of-00004.safetensors +1 -1
  5. trainer_log.jsonl +45 -0
model-00001-of-00004.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:23b002f81d758efd1ee9d063dc7de2c5502b8650262aabd1a972868d9cf6a6e4
3
  size 4877660776
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5ac2f8a3a3fabaa7944a6a64b9697f26d640453a14b893ea9fdd26417e8d492c
3
  size 4877660776
model-00002-of-00004.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:bde64166dc81980b82739f80bcab6f494d8e0158c75ffb17e19e3eae3c8d5595
3
  size 4932751008
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:da5c9de0c798d4e47a24e6eadea155fd902e7a8f97460d5f29af655b83012eed
3
  size 4932751008
model-00003-of-00004.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:34fd4f6e79255f8e81cb752a112c68e1e8c36f61033840032314aa0e0a6d14b2
3
  size 4330865200
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b1832a78151e0b6fc13be64c4748be59459b692f3f0494bda957ffb96451701b
3
  size 4330865200
model-00004-of-00004.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2417c4f78d215c9caabea424d27fbab9b65d7206d1506d72e0a5f3036a22997f
3
  size 1089994880
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:43b4543727118ffe3679f0ea570919147e2b10aea10c388c6425cdb9e870ae15
3
  size 1089994880
trainer_log.jsonl CHANGED
@@ -43,3 +43,48 @@
43
  {"current_steps": 430, "total_steps": 1314, "loss": 0.6344, "learning_rate": 5e-06, "epoch": 0.98005698005698, "percentage": 32.72, "elapsed_time": "6:16:00", "remaining_time": "12:53:00"}
44
  {"current_steps": 438, "total_steps": 1314, "eval_loss": 0.6250722408294678, "epoch": 0.9982905982905983, "percentage": 33.33, "elapsed_time": "6:30:49", "remaining_time": "13:01:39"}
45
  {"current_steps": 440, "total_steps": 1314, "loss": 0.6471, "learning_rate": 5e-06, "epoch": 1.0034188034188034, "percentage": 33.49, "elapsed_time": "6:33:11", "remaining_time": "13:01:02"}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
43
  {"current_steps": 430, "total_steps": 1314, "loss": 0.6344, "learning_rate": 5e-06, "epoch": 0.98005698005698, "percentage": 32.72, "elapsed_time": "6:16:00", "remaining_time": "12:53:00"}
44
  {"current_steps": 438, "total_steps": 1314, "eval_loss": 0.6250722408294678, "epoch": 0.9982905982905983, "percentage": 33.33, "elapsed_time": "6:30:49", "remaining_time": "13:01:39"}
45
  {"current_steps": 440, "total_steps": 1314, "loss": 0.6471, "learning_rate": 5e-06, "epoch": 1.0034188034188034, "percentage": 33.49, "elapsed_time": "6:33:11", "remaining_time": "13:01:02"}
46
+ {"current_steps": 450, "total_steps": 1314, "loss": 0.6015, "learning_rate": 5e-06, "epoch": 1.0262108262108263, "percentage": 34.25, "elapsed_time": "6:41:55", "remaining_time": "12:51:42"}
47
+ {"current_steps": 460, "total_steps": 1314, "loss": 0.5953, "learning_rate": 5e-06, "epoch": 1.049002849002849, "percentage": 35.01, "elapsed_time": "6:50:39", "remaining_time": "12:42:24"}
48
+ {"current_steps": 470, "total_steps": 1314, "loss": 0.5962, "learning_rate": 5e-06, "epoch": 1.0717948717948718, "percentage": 35.77, "elapsed_time": "6:59:24", "remaining_time": "12:33:09"}
49
+ {"current_steps": 480, "total_steps": 1314, "loss": 0.6037, "learning_rate": 5e-06, "epoch": 1.0945868945868946, "percentage": 36.53, "elapsed_time": "7:08:09", "remaining_time": "12:23:54"}
50
+ {"current_steps": 490, "total_steps": 1314, "loss": 0.6088, "learning_rate": 5e-06, "epoch": 1.1173789173789175, "percentage": 37.29, "elapsed_time": "7:16:54", "remaining_time": "12:14:42"}
51
+ {"current_steps": 500, "total_steps": 1314, "loss": 0.5987, "learning_rate": 5e-06, "epoch": 1.1401709401709401, "percentage": 38.05, "elapsed_time": "7:25:38", "remaining_time": "12:05:30"}
52
+ {"current_steps": 510, "total_steps": 1314, "loss": 0.5965, "learning_rate": 5e-06, "epoch": 1.162962962962963, "percentage": 38.81, "elapsed_time": "7:34:22", "remaining_time": "11:56:19"}
53
+ {"current_steps": 520, "total_steps": 1314, "loss": 0.6021, "learning_rate": 5e-06, "epoch": 1.1857549857549858, "percentage": 39.57, "elapsed_time": "7:43:06", "remaining_time": "11:47:08"}
54
+ {"current_steps": 530, "total_steps": 1314, "loss": 0.5952, "learning_rate": 5e-06, "epoch": 1.2085470085470085, "percentage": 40.33, "elapsed_time": "7:51:51", "remaining_time": "11:37:59"}
55
+ {"current_steps": 540, "total_steps": 1314, "loss": 0.6054, "learning_rate": 5e-06, "epoch": 1.2313390313390313, "percentage": 41.1, "elapsed_time": "8:00:35", "remaining_time": "11:28:51"}
56
+ {"current_steps": 550, "total_steps": 1314, "loss": 0.6021, "learning_rate": 5e-06, "epoch": 1.2541310541310542, "percentage": 41.86, "elapsed_time": "8:09:19", "remaining_time": "11:19:42"}
57
+ {"current_steps": 560, "total_steps": 1314, "loss": 0.6036, "learning_rate": 5e-06, "epoch": 1.2769230769230768, "percentage": 42.62, "elapsed_time": "8:18:03", "remaining_time": "11:10:36"}
58
+ {"current_steps": 570, "total_steps": 1314, "loss": 0.5976, "learning_rate": 5e-06, "epoch": 1.2997150997150997, "percentage": 43.38, "elapsed_time": "8:26:48", "remaining_time": "11:01:30"}
59
+ {"current_steps": 580, "total_steps": 1314, "loss": 0.596, "learning_rate": 5e-06, "epoch": 1.3225071225071225, "percentage": 44.14, "elapsed_time": "8:35:31", "remaining_time": "10:52:24"}
60
+ {"current_steps": 590, "total_steps": 1314, "loss": 0.5957, "learning_rate": 5e-06, "epoch": 1.3452991452991454, "percentage": 44.9, "elapsed_time": "8:44:15", "remaining_time": "10:43:19"}
61
+ {"current_steps": 600, "total_steps": 1314, "loss": 0.6014, "learning_rate": 5e-06, "epoch": 1.368091168091168, "percentage": 45.66, "elapsed_time": "8:52:59", "remaining_time": "10:34:15"}
62
+ {"current_steps": 610, "total_steps": 1314, "loss": 0.6014, "learning_rate": 5e-06, "epoch": 1.390883190883191, "percentage": 46.42, "elapsed_time": "9:01:43", "remaining_time": "10:25:12"}
63
+ {"current_steps": 620, "total_steps": 1314, "loss": 0.604, "learning_rate": 5e-06, "epoch": 1.4136752136752135, "percentage": 47.18, "elapsed_time": "9:10:27", "remaining_time": "10:16:09"}
64
+ {"current_steps": 630, "total_steps": 1314, "loss": 0.602, "learning_rate": 5e-06, "epoch": 1.4364672364672364, "percentage": 47.95, "elapsed_time": "9:19:11", "remaining_time": "10:07:06"}
65
+ {"current_steps": 640, "total_steps": 1314, "loss": 0.5969, "learning_rate": 5e-06, "epoch": 1.4592592592592593, "percentage": 48.71, "elapsed_time": "9:27:55", "remaining_time": "9:58:05"}
66
+ {"current_steps": 650, "total_steps": 1314, "loss": 0.5948, "learning_rate": 5e-06, "epoch": 1.4820512820512821, "percentage": 49.47, "elapsed_time": "9:36:39", "remaining_time": "9:49:04"}
67
+ {"current_steps": 660, "total_steps": 1314, "loss": 0.5945, "learning_rate": 5e-06, "epoch": 1.504843304843305, "percentage": 50.23, "elapsed_time": "9:45:23", "remaining_time": "9:40:04"}
68
+ {"current_steps": 670, "total_steps": 1314, "loss": 0.6022, "learning_rate": 5e-06, "epoch": 1.5276353276353276, "percentage": 50.99, "elapsed_time": "9:54:07", "remaining_time": "9:31:04"}
69
+ {"current_steps": 680, "total_steps": 1314, "loss": 0.5907, "learning_rate": 5e-06, "epoch": 1.5504273504273505, "percentage": 51.75, "elapsed_time": "10:02:50", "remaining_time": "9:22:03"}
70
+ {"current_steps": 690, "total_steps": 1314, "loss": 0.5946, "learning_rate": 5e-06, "epoch": 1.573219373219373, "percentage": 52.51, "elapsed_time": "10:11:34", "remaining_time": "9:13:04"}
71
+ {"current_steps": 700, "total_steps": 1314, "loss": 0.6101, "learning_rate": 5e-06, "epoch": 1.596011396011396, "percentage": 53.27, "elapsed_time": "10:20:17", "remaining_time": "9:04:05"}
72
+ {"current_steps": 710, "total_steps": 1314, "loss": 0.5972, "learning_rate": 5e-06, "epoch": 1.6188034188034188, "percentage": 54.03, "elapsed_time": "10:29:02", "remaining_time": "8:55:07"}
73
+ {"current_steps": 720, "total_steps": 1314, "loss": 0.6018, "learning_rate": 5e-06, "epoch": 1.6415954415954417, "percentage": 54.79, "elapsed_time": "10:37:47", "remaining_time": "8:46:10"}
74
+ {"current_steps": 730, "total_steps": 1314, "loss": 0.5949, "learning_rate": 5e-06, "epoch": 1.6643874643874645, "percentage": 55.56, "elapsed_time": "10:46:31", "remaining_time": "8:37:12"}
75
+ {"current_steps": 740, "total_steps": 1314, "loss": 0.596, "learning_rate": 5e-06, "epoch": 1.6871794871794872, "percentage": 56.32, "elapsed_time": "10:55:15", "remaining_time": "8:28:15"}
76
+ {"current_steps": 750, "total_steps": 1314, "loss": 0.5934, "learning_rate": 5e-06, "epoch": 1.7099715099715098, "percentage": 57.08, "elapsed_time": "11:03:59", "remaining_time": "8:19:19"}
77
+ {"current_steps": 760, "total_steps": 1314, "loss": 0.5973, "learning_rate": 5e-06, "epoch": 1.7327635327635327, "percentage": 57.84, "elapsed_time": "11:12:43", "remaining_time": "8:10:22"}
78
+ {"current_steps": 770, "total_steps": 1314, "loss": 0.5938, "learning_rate": 5e-06, "epoch": 1.7555555555555555, "percentage": 58.6, "elapsed_time": "11:21:27", "remaining_time": "8:01:26"}
79
+ {"current_steps": 780, "total_steps": 1314, "loss": 0.6009, "learning_rate": 5e-06, "epoch": 1.7783475783475784, "percentage": 59.36, "elapsed_time": "11:30:11", "remaining_time": "7:52:31"}
80
+ {"current_steps": 790, "total_steps": 1314, "loss": 0.5984, "learning_rate": 5e-06, "epoch": 1.8011396011396013, "percentage": 60.12, "elapsed_time": "11:38:56", "remaining_time": "7:43:36"}
81
+ {"current_steps": 800, "total_steps": 1314, "loss": 0.5853, "learning_rate": 5e-06, "epoch": 1.823931623931624, "percentage": 60.88, "elapsed_time": "11:47:40", "remaining_time": "7:34:40"}
82
+ {"current_steps": 810, "total_steps": 1314, "loss": 0.5931, "learning_rate": 5e-06, "epoch": 1.8467236467236468, "percentage": 61.64, "elapsed_time": "11:56:23", "remaining_time": "7:25:45"}
83
+ {"current_steps": 820, "total_steps": 1314, "loss": 0.5908, "learning_rate": 5e-06, "epoch": 1.8695156695156694, "percentage": 62.4, "elapsed_time": "12:05:07", "remaining_time": "7:16:50"}
84
+ {"current_steps": 830, "total_steps": 1314, "loss": 0.5946, "learning_rate": 5e-06, "epoch": 1.8923076923076922, "percentage": 63.17, "elapsed_time": "12:13:52", "remaining_time": "7:07:56"}
85
+ {"current_steps": 840, "total_steps": 1314, "loss": 0.5991, "learning_rate": 5e-06, "epoch": 1.915099715099715, "percentage": 63.93, "elapsed_time": "12:22:35", "remaining_time": "6:59:02"}
86
+ {"current_steps": 850, "total_steps": 1314, "loss": 0.5972, "learning_rate": 5e-06, "epoch": 1.937891737891738, "percentage": 64.69, "elapsed_time": "12:31:19", "remaining_time": "6:50:07"}
87
+ {"current_steps": 860, "total_steps": 1314, "loss": 0.5985, "learning_rate": 5e-06, "epoch": 1.9606837606837608, "percentage": 65.45, "elapsed_time": "12:40:03", "remaining_time": "6:41:14"}
88
+ {"current_steps": 870, "total_steps": 1314, "loss": 0.596, "learning_rate": 5e-06, "epoch": 1.9834757834757835, "percentage": 66.21, "elapsed_time": "12:48:48", "remaining_time": "6:32:21"}
89
+ {"current_steps": 877, "total_steps": 1314, "eval_loss": 0.6186895370483398, "epoch": 1.9994301994301993, "percentage": 66.74, "elapsed_time": "13:02:11", "remaining_time": "6:29:45"}
90
+ {"current_steps": 880, "total_steps": 1314, "loss": 0.6154, "learning_rate": 5e-06, "epoch": 2.006837606837607, "percentage": 66.97, "elapsed_time": "13:05:53", "remaining_time": "6:27:35"}