Hugging Face's logo

AbstractPhil
/
captionbert-8192-rotary

Feature Extraction
Transformers
TensorBoard
Safetensors
English
captionbert_structured_rotary
sentence-similarity
consensus-distillation
rotary-embeddings
custom_code
Model card Files
xet
 Metrics Community
AbstractPhil commited on
Commit
275fc76
·
verified ·
1 Parent(s): 8997526

Ship N16-seed2 (all arms retained)

Browse files
Files changed (7) hide show
  1. experiments/aafe9c59b9d696107cb4393f3a26815e84714daddb7d6eb028b5d6137b88418e/N16-seed2/best.safetensors +1 -1
  2. experiments/aafe9c59b9d696107cb4393f3a26815e84714daddb7d6eb028b5d6137b88418e/N16-seed2/eval-004500.json +16 -0
  3. experiments/aafe9c59b9d696107cb4393f3a26815e84714daddb7d6eb028b5d6137b88418e/N16-seed2/eval-005000.json +16 -0
  4. experiments/aafe9c59b9d696107cb4393f3a26815e84714daddb7d6eb028b5d6137b88418e/N16-seed2/resume.json +1 -1
  5. experiments/aafe9c59b9d696107cb4393f3a26815e84714daddb7d6eb028b5d6137b88418e/N16-seed2/state.pt +1 -1
  6. experiments/aafe9c59b9d696107cb4393f3a26815e84714daddb7d6eb028b5d6137b88418e/N16-seed2/tensorboard/events.out.tfevents.1793988039.efb5426ca2ba.52488.1 +2 -2
  7. experiments/aafe9c59b9d696107cb4393f3a26815e84714daddb7d6eb028b5d6137b88418e/N16-seed2/train.jsonl +0 -0
experiments/aafe9c59b9d696107cb4393f3a26815e84714daddb7d6eb028b5d6137b88418e/N16-seed2/best.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:db10dc1795a36e975323ecee61fcf719a77cefeb63a65b9d3342df1ef2ff5da5
3
  size 216534324
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d21ae14cd6a220dbbb372dd150196da5583de0dbad3d41fceb94dc49e9fef4af
3
  size 216534324
experiments/aafe9c59b9d696107cb4393f3a26815e84714daddb7d6eb028b5d6137b88418e/N16-seed2/eval-004500.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "above_128_tokens": {
3
+ "cos_to_target": 0.7331895232200623,
4
+ "erank": 72.62330781049799,
5
+ "n": 5000,
6
+ "self_cos": 0.011480929186906647
7
+ },
8
+ "above_128_total_rows": 29240,
9
+ "heldout": {
10
+ "cos_to_target": 0.8097149729728699,
11
+ "erank": 93.59602919647965,
12
+ "n": 5000,
13
+ "self_cos": 0.005903066371976254
14
+ },
15
+ "heldout_total_rows": 500000
16
+ }
experiments/aafe9c59b9d696107cb4393f3a26815e84714daddb7d6eb028b5d6137b88418e/N16-seed2/eval-005000.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "above_128_tokens": {
3
+ "cos_to_target": 0.7357312440872192,
4
+ "erank": 74.05763547481297,
5
+ "n": 5000,
6
+ "self_cos": 0.009572067779312262
7
+ },
8
+ "above_128_total_rows": 29240,
9
+ "heldout": {
10
+ "cos_to_target": 0.8105173110961914,
11
+ "erank": 95.59375664227889,
12
+ "n": 5000,
13
+ "self_cos": 0.004115730543474176
14
+ },
15
+ "heldout_total_rows": 500000
16
+ }
experiments/aafe9c59b9d696107cb4393f3a26815e84714daddb7d6eb028b5d6137b88418e/N16-seed2/resume.json CHANGED
@@ -3,5 +3,5 @@
3
  "config_id": "aafe9c59b9d696107cb4393f3a26815e84714daddb7d6eb028b5d6137b88418e",
4
  "dataset_id": "7adb10f8833fab44e3ebb44315b53cb65bd8010b038115d15350b18536b57e86",
5
  "seed": 2,
6
- "step": 4000
7
  }
 
3
  "config_id": "aafe9c59b9d696107cb4393f3a26815e84714daddb7d6eb028b5d6137b88418e",
4
  "dataset_id": "7adb10f8833fab44e3ebb44315b53cb65bd8010b038115d15350b18536b57e86",
5
  "seed": 2,
6
+ "step": 5300
7
  }
experiments/aafe9c59b9d696107cb4393f3a26815e84714daddb7d6eb028b5d6137b88418e/N16-seed2/state.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a55f9354a073b080b62983e418a6d78ec2d5b724f0118ae6983019e67633c046
3
  size 649636058
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:904ec0a51ab0af08f8f05c02ed170a0c0d026b8a9437fc2088a80fe261814abb
3
  size 649636058
experiments/aafe9c59b9d696107cb4393f3a26815e84714daddb7d6eb028b5d6137b88418e/N16-seed2/tensorboard/events.out.tfevents.1793988039.efb5426ca2ba.52488.1 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9d392b35db8282dbff1022695258ab5fe943012a241b4e5164a2cfe7d2732fff
3
- size 2570951
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:017b22052569131ad8914d57bd7e1c43d953cdd10c681dbee03abeb7db2d6146
3
+ size 3406415
experiments/aafe9c59b9d696107cb4393f3a26815e84714daddb7d6eb028b5d6137b88418e/N16-seed2/train.jsonl CHANGED
The diff for this file is too large to render. See raw diff