Training in progress, step 2600
Browse files- adapter_config.json +6 -6
- adapter_model.safetensors +1 -1
- angle.config +4 -4
- runs/Mar23_02-16-23_instance-20240226-074643/events.out.tfevents.1711160195.instance-20240226-074643.3613579.0 +2 -2
- runs/Mar24_10-47-38_instance-20240226-074643/events.out.tfevents.1711277263.instance-20240226-074643.300367.0 +3 -0
- training_args.bin +1 -1
adapter_config.json
CHANGED
@@ -9,7 +9,7 @@
|
|
9 |
"layers_pattern": null,
|
10 |
"layers_to_transform": null,
|
11 |
"loftq_config": {},
|
12 |
-
"lora_alpha":
|
13 |
"lora_dropout": 0.1,
|
14 |
"megatron_config": null,
|
15 |
"megatron_core": "megatron.core",
|
@@ -19,13 +19,13 @@
|
|
19 |
"rank_pattern": {},
|
20 |
"revision": null,
|
21 |
"target_modules": [
|
22 |
-
"o_proj",
|
23 |
-
"down_proj",
|
24 |
-
"up_proj",
|
25 |
-
"gate_proj",
|
26 |
"v_proj",
|
|
|
|
|
|
|
27 |
"q_proj",
|
28 |
-
"k_proj"
|
|
|
29 |
],
|
30 |
"task_type": "CAUSAL_LM",
|
31 |
"use_rslora": false
|
|
|
9 |
"layers_pattern": null,
|
10 |
"layers_to_transform": null,
|
11 |
"loftq_config": {},
|
12 |
+
"lora_alpha": 16,
|
13 |
"lora_dropout": 0.1,
|
14 |
"megatron_config": null,
|
15 |
"megatron_core": "megatron.core",
|
|
|
19 |
"rank_pattern": {},
|
20 |
"revision": null,
|
21 |
"target_modules": [
|
|
|
|
|
|
|
|
|
22 |
"v_proj",
|
23 |
+
"gate_proj",
|
24 |
+
"up_proj",
|
25 |
+
"o_proj",
|
26 |
"q_proj",
|
27 |
+
"k_proj",
|
28 |
+
"down_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:
|
3 |
size 639691872
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:da094212136e495b867fa5a2f74bf5324b02731577b79c6fe2bd4482708410e3
|
3 |
size 639691872
|
angle.config
CHANGED
@@ -6,16 +6,16 @@
|
|
6 |
"lora_config_kwargs": {
|
7 |
"task_type": "CAUSAL_LM",
|
8 |
"r": 64,
|
9 |
-
"lora_alpha":
|
10 |
"lora_dropout": 0.1,
|
11 |
"bias": "none",
|
12 |
"target_modules": [
|
13 |
-
"q_proj",
|
14 |
"v_proj",
|
15 |
-
"
|
16 |
"up_proj",
|
|
|
|
|
17 |
"k_proj",
|
18 |
-
"gate_proj",
|
19 |
"down_proj"
|
20 |
]
|
21 |
},
|
|
|
6 |
"lora_config_kwargs": {
|
7 |
"task_type": "CAUSAL_LM",
|
8 |
"r": 64,
|
9 |
+
"lora_alpha": 16,
|
10 |
"lora_dropout": 0.1,
|
11 |
"bias": "none",
|
12 |
"target_modules": [
|
|
|
13 |
"v_proj",
|
14 |
+
"gate_proj",
|
15 |
"up_proj",
|
16 |
+
"o_proj",
|
17 |
+
"q_proj",
|
18 |
"k_proj",
|
|
|
19 |
"down_proj"
|
20 |
]
|
21 |
},
|
runs/Mar23_02-16-23_instance-20240226-074643/events.out.tfevents.1711160195.instance-20240226-074643.3613579.0
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:30ff47b9b05e08bc2234b76492784b8bc20c0745b898247fb9272ef96aa21f6f
|
3 |
+
size 12941
|
runs/Mar24_10-47-38_instance-20240226-074643/events.out.tfevents.1711277263.instance-20240226-074643.300367.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d08acc73908603fe8cea89cf63c0945e253e754eced99ab1d365953c757e78ad
|
3 |
+
size 10688
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 5048
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0d25ac5d64e6b4d88ac8b861ed2a41d07744dcd0874dd58e5ddbad7c2e928dfb
|
3 |
size 5048
|