Update 4b-ad.yml
Browse files
4b-ad.yml
CHANGED
@@ -2,7 +2,7 @@ base_model: NewEden/Hamanasu-KTO-4B
|
|
2 |
model_type: AutoModelForCausalLM
|
3 |
tokenizer_type: AutoTokenizer
|
4 |
|
5 |
-
hub_model_id: NewEden/Hamanasu-4B-
|
6 |
hub_strategy: "all_checkpoints"
|
7 |
push_dataset_to_hub:
|
8 |
hf_use_auth_token: true
|
@@ -29,12 +29,6 @@ datasets:
|
|
29 |
type: dan-chat-advanced
|
30 |
- path: PocketDoc/Dans-Failuremaxx-Adventure-3
|
31 |
type: dan-chat-advanced
|
32 |
-
- path: NewEden/Misc-Mang-Sharegpt
|
33 |
-
type: dan-chat-advanced
|
34 |
-
- path: PocketDoc/Dans-Failuremaxx-Adventure
|
35 |
-
type: dan-chat-advanced
|
36 |
-
- path: PocketDoc/Dans-Failuremaxx-Adventure-2
|
37 |
-
type: dan-chat-advanced
|
38 |
- path: PocketDoc/Dans-Prosemaxx-InstructWriter-ZeroShot-2
|
39 |
type: dan-chat-advanced
|
40 |
- path: PocketDoc/Dans-Prosemaxx-InstructWriter-ZeroShot-3
|
@@ -45,14 +39,14 @@ datasets:
|
|
45 |
type: dan-chat-advanced
|
46 |
- path: PocketDoc/Dans-Prosemaxx-InstructWriter-Continue
|
47 |
type: dan-chat-advanced
|
48 |
-
- path: PocketDoc/Dans-
|
49 |
-
type: dan-chat-advanced
|
50 |
shuffle_merged_datasets: true
|
51 |
dataset_prepared_path: prepared_data
|
52 |
-
val_set_size: 0.
|
53 |
-
output_dir: ./
|
54 |
|
55 |
-
sequence_len:
|
56 |
sample_packing: true
|
57 |
pad_to_sequence_len: true
|
58 |
|
@@ -69,7 +63,7 @@ micro_batch_size: 4
|
|
69 |
num_epochs: 2
|
70 |
optimizer: adamw_bnb_8bit
|
71 |
lr_scheduler: cosine
|
72 |
-
learning_rate:
|
73 |
|
74 |
train_on_inputs: false
|
75 |
group_by_length: false
|
@@ -91,7 +85,7 @@ eval_table_size:
|
|
91 |
eval_max_new_tokens: 128
|
92 |
saves_per_epoch: 2
|
93 |
debug:
|
94 |
-
deepspeed: ./deepspeed_configs/
|
95 |
-
weight_decay: 0.
|
96 |
fsdp:
|
97 |
fsdp_config:
|
|
|
2 |
model_type: AutoModelForCausalLM
|
3 |
tokenizer_type: AutoTokenizer
|
4 |
|
5 |
+
hub_model_id: NewEden/Hamanasu-4B-Adventur-new-lr-2
|
6 |
hub_strategy: "all_checkpoints"
|
7 |
push_dataset_to_hub:
|
8 |
hf_use_auth_token: true
|
|
|
29 |
type: dan-chat-advanced
|
30 |
- path: PocketDoc/Dans-Failuremaxx-Adventure-3
|
31 |
type: dan-chat-advanced
|
|
|
|
|
|
|
|
|
|
|
|
|
32 |
- path: PocketDoc/Dans-Prosemaxx-InstructWriter-ZeroShot-2
|
33 |
type: dan-chat-advanced
|
34 |
- path: PocketDoc/Dans-Prosemaxx-InstructWriter-ZeroShot-3
|
|
|
39 |
type: dan-chat-advanced
|
40 |
- path: PocketDoc/Dans-Prosemaxx-InstructWriter-Continue
|
41 |
type: dan-chat-advanced
|
42 |
+
- path: PocketDoc/Dans-Personamaxx-VN
|
43 |
+
type: dan-chat-advanced
|
44 |
shuffle_merged_datasets: true
|
45 |
dataset_prepared_path: prepared_data
|
46 |
+
val_set_size: 0.01
|
47 |
+
output_dir: ./adventure
|
48 |
|
49 |
+
sequence_len: 32768
|
50 |
sample_packing: true
|
51 |
pad_to_sequence_len: true
|
52 |
|
|
|
63 |
num_epochs: 2
|
64 |
optimizer: adamw_bnb_8bit
|
65 |
lr_scheduler: cosine
|
66 |
+
learning_rate: 5e-6
|
67 |
|
68 |
train_on_inputs: false
|
69 |
group_by_length: false
|
|
|
85 |
eval_max_new_tokens: 128
|
86 |
saves_per_epoch: 2
|
87 |
debug:
|
88 |
+
deepspeed: ./deepspeed_configs/zero3_bf16.json
|
89 |
+
weight_decay: 0.02
|
90 |
fsdp:
|
91 |
fsdp_config:
|