Ebs's workspace
Runs
4
State
Notes
User
Tags
Created
Runtime
Sweep
batch_size
checkpointer._component_
checkpointer.checkpoint_dir
checkpointer.checkpoint_files
checkpointer.model_type
checkpointer.output_dir
dataset._component_
device
dtype
enable_activation_checkpointing
epochs
gradient_accumulation_steps
log_every_n_steps
log_peak_memory_stats
loss._component_
lr_scheduler._component_
lr_scheduler.num_warmup_steps
max_steps_per_epoch
metric_logger._component_
metric_logger.log_dir
metric_logger.name
metric_logger.project
model._component_
model.apply_lora_to_mlp
model.lora_alpha
model.lora_attn_modules
model.lora_dropout
model.lora_rank
optimizer._component_
optimizer.fused
optimizer.lr
output_dir
profiler._component_
profiler.active_steps
profiler.cpu
profiler.cuda
profiler.enabled
profiler.num_cycles
profiler.output_dir
profiler.profile_memory
profiler.record_shapes
profiler.wait_steps
profiler.warmup_steps
profiler.with_flops
Finished
-
ebs
1m 20s
-
16
torchtune.training.FullModelHFCheckpointer
/tmp/Qwen2-1.5B-Instruct
["model.safetensors"]
QWEN2
/tmp/Qwen2-1.5B-Instruct-lora-finetune
torchtune.datasets.alpaca_cleaned_dataset
cuda
bf16
true
1
1
1
true
torchtune.modules.loss.CEWithChunkedOutputLoss
torchtune.training.lr_schedulers.get_cosine_schedule_with_warmup
100
100
torchtune.training.metric_logging.WandBLogger
/tmp/Qwen2-1.5B-Instruct-lora-finetune
baseline-bsz16-ga1
qwen2-1.5b-lora-distributed
torchtune.models.qwen2.lora_qwen2_1_5b
false
64
["q_proj","v_proj"]
0
32
torch.optim.AdamW
true
0.00002
/tmp/Qwen2-1.5B-Instruct-lora-finetune
torchtune.training.setup_torch_profiler
2
true
true
false
1
/tmp/Qwen2-1.5B-Instruct-lora-finetune/profiling_outputs
false
true
5
5
false
Finished
-
ebs
8m 33s
-
1
torchtune.training.FullModelHFCheckpointer
/tmp/Qwen2-1.5B-Instruct
["model.safetensors"]
QWEN2
/tmp/Qwen2-1.5B-Instruct-lora-finetune
torchtune.datasets.alpaca_cleaned_dataset
cuda
bf16
true
1
16
1
true
torchtune.modules.loss.CEWithChunkedOutputLoss
torchtune.training.lr_schedulers.get_cosine_schedule_with_warmup
100
100
torchtune.training.metric_logging.WandBLogger
/tmp/Qwen2-1.5B-Instruct-lora-finetune
baseline-bsz1-ga16
qwen2-1.5b-lora-distributed
torchtune.models.qwen2.lora_qwen2_1_5b
false
64
["q_proj","v_proj"]
0
32
torch.optim.AdamW
true
0.00002
/tmp/Qwen2-1.5B-Instruct-lora-finetune
torchtune.training.setup_torch_profiler
2
true
true
false
1
/tmp/Qwen2-1.5B-Instruct-lora-finetune/profiling_outputs
false
true
5
5
false
Finished
-
ebs
1m 21s
-
16
torchtune.training.FullModelHFCheckpointer
/tmp/Qwen2-1.5B-Instruct
["model.safetensors"]
QWEN2
/tmp/Qwen2-1.5B-Instruct-lora-finetune
torchtune.datasets.alpaca_cleaned_dataset
cuda
bf16
true
1
1
1
true
torchtune.modules.loss.CEWithChunkedOutputLoss
torchtune.training.lr_schedulers.get_cosine_schedule_with_warmup
100
100
torchtune.training.metric_logging.WandBLogger
/tmp/Qwen2-1.5B-Instruct-lora-finetune
new-bsz16-ga1
qwen2-1.5b-lora-distributed
torchtune.models.qwen2.lora_qwen2_1_5b
false
64
["q_proj","v_proj"]
0
32
torch.optim.AdamW
true
0.00002
/tmp/Qwen2-1.5B-Instruct-lora-finetune
torchtune.training.setup_torch_profiler
2
true
true
false
1
/tmp/Qwen2-1.5B-Instruct-lora-finetune/profiling_outputs
false
true
5
5
false
Finished
-
ebs
8m 34s
-
1
torchtune.training.FullModelHFCheckpointer
/tmp/Qwen2-1.5B-Instruct
["model.safetensors"]
QWEN2
/tmp/Qwen2-1.5B-Instruct-lora-finetune
torchtune.datasets.alpaca_cleaned_dataset
cuda
bf16
true
1
16
1
true
torchtune.modules.loss.CEWithChunkedOutputLoss
torchtune.training.lr_schedulers.get_cosine_schedule_with_warmup
100
100
torchtune.training.metric_logging.WandBLogger
/tmp/Qwen2-1.5B-Instruct-lora-finetune
new-bsz1-ga16
qwen2-1.5b-lora-distributed
torchtune.models.qwen2.lora_qwen2_1_5b
false
64
["q_proj","v_proj"]
0
32
torch.optim.AdamW
true
0.00002
/tmp/Qwen2-1.5B-Instruct-lora-finetune
torchtune.training.setup_torch_profiler
2
true
true
false
1
/tmp/Qwen2-1.5B-Instruct-lora-finetune/profiling_outputs
false
true
5
5
false
1-4
of 4