Hugging Face's logo

SeanLee97
/
bellm-llama-7b-nli

PEFT
TensorBoard
Safetensors
Model card Files Metrics Community
SeanLee97 commited on
Commit
c4e2359
•
1 Parent(s): 559d016

Training in progress, step 200

Browse files
Files changed (28) hide show
  1. adapter_config.json +8 -3
  2. adapter_model.safetensors +2 -2
  3. angle.config +12 -3
  4. best-checkpoint/adapter_config.json +7 -2
  5. best-checkpoint/adapter_model.safetensors +2 -2
  6. runs/Apr08_00-39-31_instance-20240226-074643/events.out.tfevents.1712536780.instance-20240226-074643.1438858.0 +2 -2
  7. runs/Apr08_06-42-53_instance-20240226-074643/events.out.tfevents.1712558587.instance-20240226-074643.1602513.0 +3 -0
  8. runs/Apr08_06-49-36_instance-20240226-074643/events.out.tfevents.1712558994.instance-20240226-074643.1603332.0 +3 -0
  9. runs/Apr08_06-51-20_instance-20240226-074643/events.out.tfevents.1712559094.instance-20240226-074643.1615047.0 +3 -0
  10. runs/Apr08_06-53-09_instance-20240226-074643/events.out.tfevents.1712559203.instance-20240226-074643.1616035.0 +3 -0
  11. runs/Apr08_06-55-24_instance-20240226-074643/events.out.tfevents.1712559338.instance-20240226-074643.1616883.0 +3 -0
  12. runs/Apr08_06-56-51_instance-20240226-074643/events.out.tfevents.1712559424.instance-20240226-074643.1617669.0 +3 -0
  13. runs/Apr08_06-58-25_instance-20240226-074643/events.out.tfevents.1712559515.instance-20240226-074643.1618512.0 +3 -0
  14. runs/Apr08_07-00-41_instance-20240226-074643/events.out.tfevents.1712559651.instance-20240226-074643.1619413.0 +3 -0
  15. runs/Apr08_07-02-48_instance-20240226-074643/events.out.tfevents.1712559782.instance-20240226-074643.1620199.0 +3 -0
  16. runs/Apr08_07-04-54_instance-20240226-074643/events.out.tfevents.1712559908.instance-20240226-074643.1621099.0 +3 -0
  17. runs/Apr08_07-06-10_instance-20240226-074643/events.out.tfevents.1712559984.instance-20240226-074643.1621815.0 +3 -0
  18. runs/Apr08_07-08-25_instance-20240226-074643/events.out.tfevents.1712560115.instance-20240226-074643.1622642.0 +3 -0
  19. runs/Apr08_07-09-20_instance-20240226-074643/events.out.tfevents.1712560170.instance-20240226-074643.1623332.0 +3 -0
  20. runs/Apr08_07-12-28_instance-20240226-074643/events.out.tfevents.1712560354.instance-20240226-074643.1624318.0 +3 -0
  21. runs/Apr08_07-14-12_instance-20240226-074643/events.out.tfevents.1712560459.instance-20240226-074643.1625120.0 +3 -0
  22. runs/Apr08_07-15-58_instance-20240226-074643/events.out.tfevents.1712560564.instance-20240226-074643.1625943.0 +3 -0
  23. runs/Apr08_07-17-31_instance-20240226-074643/events.out.tfevents.1712560657.instance-20240226-074643.1626685.0 +3 -0
  24. runs/Apr08_07-19-59_instance-20240226-074643/events.out.tfevents.1712560805.instance-20240226-074643.1637741.0 +3 -0
  25. runs/Apr08_07-21-28_instance-20240226-074643/events.out.tfevents.1712560893.instance-20240226-074643.1638487.0 +3 -0
  26. runs/Apr08_07-28-25_instance-20240226-074643/events.out.tfevents.1712561314.instance-20240226-074643.1639309.0 +3 -0
  27. runs/Apr08_07-50-27_instance-20240226-074643/events.out.tfevents.1712562633.instance-20240226-074643.1653113.0 +3 -0
  28. training_args.bin +1 -1
adapter_config.json CHANGED
@@ -10,7 +10,7 @@
10
  "layers_to_transform": null,
11
  "loftq_config": {},
12
  "lora_alpha": 32,
13
- "lora_dropout": 0.05,
14
  "megatron_config": null,
15
  "megatron_core": "megatron.core",
16
  "modules_to_save": null,
@@ -19,8 +19,13 @@
19
  "rank_pattern": {},
20
  "revision": null,
21
  "target_modules": [
22
- "v_proj",
23
- "q_proj"
 
 
 
 
 
24
  ],
25
  "task_type": "CAUSAL_LM",
26
  "use_rslora": false
 
10
  "layers_to_transform": null,
11
  "loftq_config": {},
12
  "lora_alpha": 32,
13
+ "lora_dropout": 0.1,
14
  "megatron_config": null,
15
  "megatron_core": "megatron.core",
16
  "modules_to_save": null,
 
19
  "rank_pattern": {},
20
  "revision": null,
21
  "target_modules": [
22
+ "k_proj",
23
+ "o_proj",
24
+ "up_proj",
25
+ "down_proj",
26
+ "q_proj",
27
+ "gate_proj",
28
+ "v_proj"
29
  ],
30
  "task_type": "CAUSAL_LM",
31
  "use_rslora": false
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:4bf3b5ef82f4f93bd5f23d1c930ea1cca50a8e00a6b739235e44c32c14e482bf
3
- size 33571752
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cf3647c08d4a8b4b4e08744593df757d857d7c7425d855761e47fd5573a81ad5
3
+ size 319876032
angle.config CHANGED
@@ -1,14 +1,23 @@
1
  {
2
  "model_name_or_path": "NousResearch/Llama-2-7b-hf",
3
- "max_length": 75,
4
  "model_kwargs": {},
5
  "pooling_strategy": "cls",
6
  "lora_config_kwargs": {
7
  "task_type": "CAUSAL_LM",
8
  "r": 32,
9
  "lora_alpha": 32,
10
- "lora_dropout": 0.05,
11
- "bias": "none"
 
 
 
 
 
 
 
 
 
12
  },
13
  "apply_lora": 1
14
  }
 
1
  {
2
  "model_name_or_path": "NousResearch/Llama-2-7b-hf",
3
+ "max_length": 50,
4
  "model_kwargs": {},
5
  "pooling_strategy": "cls",
6
  "lora_config_kwargs": {
7
  "task_type": "CAUSAL_LM",
8
  "r": 32,
9
  "lora_alpha": 32,
10
+ "lora_dropout": 0.1,
11
+ "bias": "none",
12
+ "target_modules": [
13
+ "q_proj",
14
+ "k_proj",
15
+ "up_proj",
16
+ "down_proj",
17
+ "v_proj",
18
+ "gate_proj",
19
+ "o_proj"
20
+ ]
21
  },
22
  "apply_lora": 1
23
  }
best-checkpoint/adapter_config.json CHANGED
@@ -10,7 +10,7 @@
10
  "layers_to_transform": null,
11
  "loftq_config": {},
12
  "lora_alpha": 32,
13
- "lora_dropout": 0.05,
14
  "megatron_config": null,
15
  "megatron_core": "megatron.core",
16
  "modules_to_save": null,
@@ -19,8 +19,13 @@
19
  "rank_pattern": {},
20
  "revision": null,
21
  "target_modules": [
 
 
 
 
22
  "q_proj",
23
- "v_proj"
 
24
  ],
25
  "task_type": "CAUSAL_LM",
26
  "use_rslora": false
 
10
  "layers_to_transform": null,
11
  "loftq_config": {},
12
  "lora_alpha": 32,
13
+ "lora_dropout": 0.1,
14
  "megatron_config": null,
15
  "megatron_core": "megatron.core",
16
  "modules_to_save": null,
 
19
  "rank_pattern": {},
20
  "revision": null,
21
  "target_modules": [
22
+ "k_proj",
23
+ "down_proj",
24
+ "v_proj",
25
+ "gate_proj",
26
  "q_proj",
27
+ "o_proj",
28
+ "up_proj"
29
  ],
30
  "task_type": "CAUSAL_LM",
31
  "use_rslora": false
best-checkpoint/adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ee901b2a75321a767f8ffab07ff5b417854aef9755d37ff31472b1237395df25
3
- size 33571752
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b00bb4a18247350389502d8666fc7443b98139abb8dd6954431732016bbf78d
3
+ size 319876032
runs/Apr08_00-39-31_instance-20240226-074643/events.out.tfevents.1712536780.instance-20240226-074643.1438858.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c827e0bc3f963bae09c043539188a8cbfa8ff1b71cb6a81034ea7c4b187f0450
3
- size 94045
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0348a269278e9ef1060d338a0ba8ba37a86a0cef50fd3c807ad6d36505fb4b54
3
+ size 94260
runs/Apr08_06-42-53_instance-20240226-074643/events.out.tfevents.1712558587.instance-20240226-074643.1602513.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bc59d80f15041e6a4a9a06f99b8f7cb9aed400c5ca67d6c108b072d3fac91704
3
+ size 4830
runs/Apr08_06-49-36_instance-20240226-074643/events.out.tfevents.1712558994.instance-20240226-074643.1603332.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7a67a1f32005636ed6c86a6038726ed516e6305a61edb2af3b10e1524b8784e0
3
+ size 4830
runs/Apr08_06-51-20_instance-20240226-074643/events.out.tfevents.1712559094.instance-20240226-074643.1615047.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:844c28e762c50e3743f95966def00c6524cce9118ff7dd542044258e82b619b8
3
+ size 4830
runs/Apr08_06-53-09_instance-20240226-074643/events.out.tfevents.1712559203.instance-20240226-074643.1616035.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:312863573d3627ea08ed72ae0db78310b456860b506b686b3197e083293c4a06
3
+ size 4830
runs/Apr08_06-55-24_instance-20240226-074643/events.out.tfevents.1712559338.instance-20240226-074643.1616883.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3bfbf4b422999dad01691dd1fda27bdeaf9200527d2d0d7ddb92c0a9cb829088
3
+ size 4830
runs/Apr08_06-56-51_instance-20240226-074643/events.out.tfevents.1712559424.instance-20240226-074643.1617669.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d674d6a8169958a24d437515ee7a4a5699974b7362d3166bc09d7605e431fb59
3
+ size 4829
runs/Apr08_06-58-25_instance-20240226-074643/events.out.tfevents.1712559515.instance-20240226-074643.1618512.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:41b69a2b202f91904661a595dbade33d2418509de34c0042a6eedc9428b0255b
3
+ size 4829
runs/Apr08_07-00-41_instance-20240226-074643/events.out.tfevents.1712559651.instance-20240226-074643.1619413.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5102574742da68b1aab8b268018b7d94b8de6ba32e2e6dae1de042f29aafa863
3
+ size 4829
runs/Apr08_07-02-48_instance-20240226-074643/events.out.tfevents.1712559782.instance-20240226-074643.1620199.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:738e6151cb6630e0d6c5f5052f0c63534013c82750863488e4c72ca70046f07f
3
+ size 4830
runs/Apr08_07-04-54_instance-20240226-074643/events.out.tfevents.1712559908.instance-20240226-074643.1621099.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c82a17f5cb8bd2a47bf607b805977503d24c43675572ff35ce6272a7a60ca369
3
+ size 4830
runs/Apr08_07-06-10_instance-20240226-074643/events.out.tfevents.1712559984.instance-20240226-074643.1621815.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8d4110f34a48559dd13919b6dd358a8f9489d7ccc999d6d5ca25fd6583da6e7a
3
+ size 4830
runs/Apr08_07-08-25_instance-20240226-074643/events.out.tfevents.1712560115.instance-20240226-074643.1622642.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0ef2ff59fc16fff6ed92d5c0556f68b95e261d99efc5d764ee125b0aaa100149
3
+ size 4830
runs/Apr08_07-09-20_instance-20240226-074643/events.out.tfevents.1712560170.instance-20240226-074643.1623332.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2c18275d8d714698b3a34f28998c4d4bd99c6a1dce2631f468e452ee61bdd22e
3
+ size 5178
runs/Apr08_07-12-28_instance-20240226-074643/events.out.tfevents.1712560354.instance-20240226-074643.1624318.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:821fd2c21a4d30584957c5cbf47d3c46c153c14a9af6484ffbf8c1ebaaf20ddb
3
+ size 5278
runs/Apr08_07-14-12_instance-20240226-074643/events.out.tfevents.1712560459.instance-20240226-074643.1625120.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3e1d8aa97b77b9e18395a8d24b47be247acf3267c61f31653af77adb0b7ba330
3
+ size 5278
runs/Apr08_07-15-58_instance-20240226-074643/events.out.tfevents.1712560564.instance-20240226-074643.1625943.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6df284d1fcac458ea2ff882ace7970d679c99b9ffd396683c6ce71b531d6148a
3
+ size 5278
runs/Apr08_07-17-31_instance-20240226-074643/events.out.tfevents.1712560657.instance-20240226-074643.1626685.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:82de84a2040832b1229d92757031194ffab7e20c77caec0c3715d69513966b68
3
+ size 5626
runs/Apr08_07-19-59_instance-20240226-074643/events.out.tfevents.1712560805.instance-20240226-074643.1637741.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e5fc4fced2b10a37d7975f89048a067b622e9a1085dfb54789ce067e7d4c2a22
3
+ size 5278
runs/Apr08_07-21-28_instance-20240226-074643/events.out.tfevents.1712560893.instance-20240226-074643.1638487.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:588119306a7f16f3f687381da088a2ec196a2caee06125c6543163cb4f51637f
3
+ size 5279
runs/Apr08_07-28-25_instance-20240226-074643/events.out.tfevents.1712561314.instance-20240226-074643.1639309.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f867cebd6aab7939f7ba8fe5337be1c81446ef99968439ecf42793296f4d202b
3
+ size 5279
runs/Apr08_07-50-27_instance-20240226-074643/events.out.tfevents.1712562633.instance-20240226-074643.1653113.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5909b318587e318df4ebbe436d8f79515a91eaf76d0f3a86164f69358e9fa0b4
3
+ size 5697
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:4e96294bc71998a1bc830d8c266b4cf14ca872ad951992e63f71f2ca616537fd
3
  size 5048
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9ba19f97ec4cbca4fef18206ec5671b2cd242c7abd8098136a4b5f9896d91766
3
  size 5048