File size: 2,018 Bytes
77fd4d0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
base_model: cognitivecomputations/dolphin-2.9-llama3-8b
model_type: LlamaForCausalLM
tokenizer_type: AutoTokenizer  # PreTrainedTokenizerFast

load_in_8bit: false
load_in_4bit: true
strict: false

datasets:
  - path: taddeusb90/finbro-v0.1.0
    type: alpaca
dataset_prepared_path: last_run_prepared
val_set_size: 0.01
output_dir: ./out/finbro-v0.1.0-dolphin-2.9-llama-3-8B-instruct-1m

adapter: qlora
lora_model_dir:

sequence_len: 8192
sample_packing: false
pad_to_sequence_len: true

lora_r: 8
lora_alpha: 16
lora_dropout: 0.05
lora_target_modules:
lora_target_linear: true
lora_fan_in_fan_out:

wandb_project: finbro-v0.1.0-dolphin-2.9-llama-3-8B-instruct
wandb_entity: sigmance
wandb_watch: "true"
wandb_name: finbro-v0.1.0-dolphin-2.9-llama-3-8B-instruct-1m
wandb_log_model: "true"

use_pose: true
pose_max_context_len: 1048576

# lora_on_cpu: 
overrides_of_model_config:
  rope_theta: 500000.0
  max_position_embeddings: 1048576
  rope_scaling: 

gradient_accumulation_steps: 4
micro_batch_size: 1
num_epochs: 4
optimizer: adamw_torch
lr_scheduler: cosine
learning_rate: 0.0002

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false

gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: true
early_stopping_patience: 50
resume_from_checkpoint: ./out/finbro-v0.1.0-dolphin-2.9-llama-3-8B-instruct-1m/checkpoint-8800
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true


warmup_steps: 1
save_steps: 100
eval_steps: 100
# evals_per_epoch: 4
eval_table_size:
# saves_per_epoch: 1
debug:
deepspeed: 
weight_decay: 0.0
fsdp:
  - full_shard
  - auto_wrap
fsdp_config:
  fsdp_limit_all_gathers: true
  fsdp_sync_module_states: true
  fsdp_offload_params: true
  fsdp_use_orig_params: false
  fsdp_cpu_ram_efficient_loading: true
  fsdp_auto_wrap_policy: TRANSFORMER_BASED_WRAP
  fsdp_transformer_layer_cls_to_wrap: LlamaDecoderLayer
  fsdp_state_dict_type: FULL_STATE_DICT
  fsdp_sharding_strategy: FULL_SHARD
special_tokens:
  pad_token: <|end_of_text|>