cryptography / training_config.yml
hfcsrd's picture
Upload training_config.yml with huggingface_hub
81fb5e6 verified
raw
history blame
2.07 kB
model:
_component_: models.lora_mmllama3_8b
lora_attn_modules:
- q_proj
- v_proj
apply_lora_to_mlp: false
apply_lora_to_output: false
lora_rank: 16
lora_alpha: 32
perception_tokens: 2
use_clip: false
tokenizer:
_component_: models.a2a_tokenizer
path: models/tokenizer.model
checkpointer:
_component_: torchtune.utils.FullModelMetaCheckpointer
checkpoint_dir: /workspace/omega_a2a/training
checkpoint_files:
- consolidated.00.pth
adapter_checkpoint: null
recipe_checkpoint: null
output_dir: /workspace/omega_a2a/checkpoints
model_type: LLAMA3
resume_from_checkpoint: false
interim_checkpoint_steps: 5000
interim_gen_steps: null
max_new_tokens: 100
temperature: 0.8
top_k: 230
dataset:
_component_: ds.EvenBatcher
buffer_size: 36
dataset:
_component_: ds.RoundRobinDataset
datasets:
- _component_: ds.OmegaVideoCaptionDataset
length: 500000
- _component_: ds.LlavaInstructDataset
dataset_path: ds/coco_llava_instruct/output.parquet
train_on_input: false
- _component_: ds.LlavaInstructDataset
dataset_path: ds/vision_flan/output.parquet
train_on_input: false
- _component_: ds.CaptionInstructDataset
dataset_path: ds/sam_llava/output.parquet
train_on_input: false
seed: null
shuffle: true
batch_size: 4
optimizer:
_component_: torch.optim.AdamW
weight_decay: 0.0001
lr: 3.0e-05
lr_scheduler:
_component_: torchtune.modules.get_cosine_schedule_with_warmup
num_warmup_steps: 100
loss:
_component_: torch.nn.CrossEntropyLoss
epochs: 6
max_steps_per_epoch: null
gradient_accumulation_steps: 64
compile: false
output_dir: /tmp/lora_finetune_output
metric_logger:
_component_: torchtune.utils.metric_logging.DiskLogger
log_dir: ${output_dir}
log_every_n_steps: null
device: cuda
dtype: bf16
enable_activation_checkpointing: false
profiler:
_component_: torchtune.utils.profiler
enabled: false
inference:
prompt_template: 'Video:
{video}
Caption the previous video.'
max_new_tokens: 100
temperature: 0.6
top_k: 170
quantizer: null