Model_name
string | Train_size
int64 | Test_size
int64 | arg
dict | lora
list | Parameters
int64 | Trainable_parameters
int64 | r
int64 | Memory Allocation
string | Training Time
string | accuracy
float64 | f1_macro
float64 | f1_weighted
float64 | precision
float64 | recall
float64 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Qwen/Qwen2-1.5B
| 50,775
| 12,652
|
{
"adafactor": false,
"adam_beta1": 0.9,
"adam_beta2": 0.999,
"adam_epsilon": 1e-8,
"bf16": false,
"fp16": false,
"fp16_opt_level": "O1",
"gradient_accumulation_steps": 4,
"half_precision_backend": "auto",
"label_smoothing_factor": 0,
"learning_rate": 0.00005,
"lr_scheduler_type": "linear",
"max_grad_norm": 1,
"max_steps": -1,
"n_gpu": 2,
"num_train_epochs": 1,
"optim": "adamw_8bit",
"optim_args": "Not have",
"per_device_eval_batch_size": 8,
"per_device_train_batch_size": 8,
"warmup_ratio": 0,
"warmup_steps": 5,
"weight_decay": 0.01
}
|
[
"down_proj",
"gate_proj",
"o_proj",
"up_proj"
] | 1,551,050,240
| 7,761,408
| 8
|
7436.54
|
2985.49
| 0.887053
| 0.880311
| 0.887285
| 0.881419
| 0.879712
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.