Igoro's group workspace
Group: P3_2jmuwy18
State
Notes
User
Tags
Created
Runtime
Sweep
activation
adlr_autoresume
adlr_autoresume_interval
apply_query_key_layer_scaling
attention-dropout
attention_config
attention_dropout
attention_softmax_in_fp32
batch_size
bias-gelu-fusion
bias_dropout_fusion
bias_gelu_fusion
char_level_ppl
checkpoint-activations
checkpoint-num-layers
checkpoint_activations
checkpoint_factor
checkpoint_in_cpu
checkpoint_num_layers
checkpoint_scale
checkpoint_validation_with_forward_pass
clip_grad
config_files.20B.yml
config_files.20B_P3.yml
config_files.20B_code.yml
config_files.20B_fork.yml
config_files.20B_math.yml
config_files.test20B.yml
contiguous_checkpointing
coord_check
curriculum_seqlen
data-impl
data-path
data_impl
data_path
deepscale
deepspeed
deepspeed_activation_checkpointing
deepspeed_mpi
deepspeed_slurm
detect_nvlink_pairs
distributed-backend
distributed_backend
dump_state
Finished
stellaathena
9h 57m 53s
-
gelu
false
1000
false
-
["global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global"]
0
false
4
-
false
true
false
-
-
true
-
false
1
-
false
1
-
# GPT-2 pretraining setup
{
# parallelism settings ( you will want to change these based on your cluster setup, ideally scheduling pipeline stages
# across the node boundaries )
"pipe-parallel-size": 4,
"model-parallel-size": 2,
# model settings
"num-layers": 44,
"hidden-size": 6144,
"num-attention-heads": 64,
"seq-length": 2048,
"max-position-embeddings": 2048,
"norm": "layernorm",
"pos-emb": "rotary",
"rotary_pct": 0.25,
"no-weight-tying": true,
"gpt_j_residual": true,
"output_layer_parallelism": "column",
# these should provide some speedup but takes a while to build, set to true if desired
"scaled-upper-triang-masked-softmax-fusion": false,
"bias-gelu-fusion": true,
# init methods
"init_method": "small_init",
"output_layer_init_method": "wang_init",
# optimizer settings
"optimizer": {
"type": "Adam",
"params": {
"lr": 9.7e-8,
"betas": [0.9, 0.95],
"eps": 1.0e-8,
}
},
"min_lr": 9.7e-8,
"override-lr-scheduler": True,
"zero_optimization": {
"stage": 1,
"allgather_partitions": True,
"allgather_bucket_size": 1260000000,
"overlap_comm": True,
"reduce_scatter": True,
"reduce_bucket_size": 1260000000,
"contiguous_gradients": True,
"cpu_offload": False
},
# batch / data settings (assuming 32 GPUs)
"train_micro_batch_size_per_gpu": 4,
"gradient_accumulation_steps": 16,
"data-impl": "mmap",
# activation checkpointing
"checkpoint-activations": true,
"checkpoint-num-layers": 1,
"partition-activations": false,
"synchronize-each-layer": true,
# regularization
"gradient_clipping": 1.0,
"weight-decay": 0.01,
"hidden-dropout": 0,
"attention-dropout": 0,
# precision settings
"fp16": {
"fp16": true,
"enabled": true,
"loss_scale": 0,
"loss_scale_window": 1000,
"initial_scale_power": 12,
"hysteresis": 2,
"min_loss_scale": 1
},
# misc. training settings
"train-iters": 200000,
"distributed-backend": "nccl",
"lr-decay-style": "constant",
"save-interval": 500,
"eval-interval": 100,
"eval-iters": 1,
# "eval_tasks": ["lambada", "piqa", "hellaswag", "winogrande", "mathqa", "pubmedqa"],
# logging
"log-interval": 2,
"steps_per_print": 2,
"wall_clock_breakdown": false,
### NEW DATA: ####
"tokenizer_type": "HFTokenizer",
"vocab-file": "/mnt/ssd-1/data/20B_tokenizer.json",
"train_data_paths": ["/mnt/ssd-1/P3/P3_combined/train_text_document"],
"valid_data_paths": ["/mnt/ssd-1/P3/P3_combined/validation_text_document"],
"test_data_paths": ["/mnt/ssd-1/P3/P3_combined/test_text_document"],
"train_data_weights": [1.0], "valid_data_weights": [1.0], "test_data_weights": [1.0],
"save": "/mnt/ssd-1/20B_P3",
"load": "/mnt/ssd-1/20B_P3",
"keep_last_n_checkpoints": 3,
"tensorboard-dir": "/mnt/ssd-1/tensorboard",
"log-dir": "/mnt/ssd-1/logs",
"wandb_team": "eleutherai",
"wandb_project": "gpt-thicc",
"wandb_group": "P3"
}
-
-
-
-
false
-
-
-
-
mmap
-
false
true
true
false
-
false
-
nccl
false
Finished
stellaathena
9h 57m 56s
-
gelu
false
1000
false
-
["global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global"]
0
false
4
-
false
true
false
-
-
true
-
false
1
-
false
1
-
# GPT-2 pretraining setup
{
# parallelism settings ( you will want to change these based on your cluster setup, ideally scheduling pipeline stages
# across the node boundaries )
"pipe-parallel-size": 4,
"model-parallel-size": 2,
# model settings
"num-layers": 44,
"hidden-size": 6144,
"num-attention-heads": 64,
"seq-length": 2048,
"max-position-embeddings": 2048,
"norm": "layernorm",
"pos-emb": "rotary",
"rotary_pct": 0.25,
"no-weight-tying": true,
"gpt_j_residual": true,
"output_layer_parallelism": "column",
# these should provide some speedup but takes a while to build, set to true if desired
"scaled-upper-triang-masked-softmax-fusion": false,
"bias-gelu-fusion": true,
# init methods
"init_method": "small_init",
"output_layer_init_method": "wang_init",
# optimizer settings
"optimizer": {
"type": "Adam",
"params": {
"lr": 9.7e-8,
"betas": [0.9, 0.95],
"eps": 1.0e-8,
}
},
"min_lr": 9.7e-8,
"override-lr-scheduler": True,
"zero_optimization": {
"stage": 1,
"allgather_partitions": True,
"allgather_bucket_size": 1260000000,
"overlap_comm": True,
"reduce_scatter": True,
"reduce_bucket_size": 1260000000,
"contiguous_gradients": True,
"cpu_offload": False
},
# batch / data settings (assuming 32 GPUs)
"train_micro_batch_size_per_gpu": 4,
"gradient_accumulation_steps": 16,
"data-impl": "mmap",
# activation checkpointing
"checkpoint-activations": true,
"checkpoint-num-layers": 1,
"partition-activations": false,
"synchronize-each-layer": true,
# regularization
"gradient_clipping": 1.0,
"weight-decay": 0.01,
"hidden-dropout": 0,
"attention-dropout": 0,
# precision settings
"fp16": {
"fp16": true,
"enabled": true,
"loss_scale": 0,
"loss_scale_window": 1000,
"initial_scale_power": 12,
"hysteresis": 2,
"min_loss_scale": 1
},
# misc. training settings
"train-iters": 200000,
"distributed-backend": "nccl",
"lr-decay-style": "constant",
"save-interval": 500,
"eval-interval": 100,
"eval-iters": 1,
# "eval_tasks": ["lambada", "piqa", "hellaswag", "winogrande", "mathqa", "pubmedqa"],
# logging
"log-interval": 2,
"steps_per_print": 2,
"wall_clock_breakdown": false,
### NEW DATA: ####
"tokenizer_type": "HFTokenizer",
"vocab-file": "/mnt/ssd-1/data/20B_tokenizer.json",
"train_data_paths": ["/mnt/ssd-1/P3/P3_combined/train_text_document"],
"valid_data_paths": ["/mnt/ssd-1/P3/P3_combined/validation_text_document"],
"test_data_paths": ["/mnt/ssd-1/P3/P3_combined/test_text_document"],
"train_data_weights": [1.0], "valid_data_weights": [1.0], "test_data_weights": [1.0],
"save": "/mnt/ssd-1/20B_P3",
"load": "/mnt/ssd-1/20B_P3",
"keep_last_n_checkpoints": 3,
"tensorboard-dir": "/mnt/ssd-1/tensorboard",
"log-dir": "/mnt/ssd-1/logs",
"wandb_team": "eleutherai",
"wandb_project": "gpt-thicc",
"wandb_group": "P3"
}
-
-
-
-
false
-
-
-
-
mmap
-
false
true
true
false
-
false
-
nccl
false
Finished
stellaathena
9h 57m 51s
-
gelu
false
1000
false
-
["global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global"]
0
false
4
-
false
true
false
-
-
true
-
false
1
-
false
1
-
# GPT-2 pretraining setup
{
# parallelism settings ( you will want to change these based on your cluster setup, ideally scheduling pipeline stages
# across the node boundaries )
"pipe-parallel-size": 4,
"model-parallel-size": 2,
# model settings
"num-layers": 44,
"hidden-size": 6144,
"num-attention-heads": 64,
"seq-length": 2048,
"max-position-embeddings": 2048,
"norm": "layernorm",
"pos-emb": "rotary",
"rotary_pct": 0.25,
"no-weight-tying": true,
"gpt_j_residual": true,
"output_layer_parallelism": "column",
# these should provide some speedup but takes a while to build, set to true if desired
"scaled-upper-triang-masked-softmax-fusion": false,
"bias-gelu-fusion": true,
# init methods
"init_method": "small_init",
"output_layer_init_method": "wang_init",
# optimizer settings
"optimizer": {
"type": "Adam",
"params": {
"lr": 9.7e-8,
"betas": [0.9, 0.95],
"eps": 1.0e-8,
}
},
"min_lr": 9.7e-8,
"override-lr-scheduler": True,
"zero_optimization": {
"stage": 1,
"allgather_partitions": True,
"allgather_bucket_size": 1260000000,
"overlap_comm": True,
"reduce_scatter": True,
"reduce_bucket_size": 1260000000,
"contiguous_gradients": True,
"cpu_offload": False
},
# batch / data settings (assuming 32 GPUs)
"train_micro_batch_size_per_gpu": 4,
"gradient_accumulation_steps": 16,
"data-impl": "mmap",
# activation checkpointing
"checkpoint-activations": true,
"checkpoint-num-layers": 1,
"partition-activations": false,
"synchronize-each-layer": true,
# regularization
"gradient_clipping": 1.0,
"weight-decay": 0.01,
"hidden-dropout": 0,
"attention-dropout": 0,
# precision settings
"fp16": {
"fp16": true,
"enabled": true,
"loss_scale": 0,
"loss_scale_window": 1000,
"initial_scale_power": 12,
"hysteresis": 2,
"min_loss_scale": 1
},
# misc. training settings
"train-iters": 200000,
"distributed-backend": "nccl",
"lr-decay-style": "constant",
"save-interval": 500,
"eval-interval": 100,
"eval-iters": 1,
# "eval_tasks": ["lambada", "piqa", "hellaswag", "winogrande", "mathqa", "pubmedqa"],
# logging
"log-interval": 2,
"steps_per_print": 2,
"wall_clock_breakdown": false,
### NEW DATA: ####
"tokenizer_type": "HFTokenizer",
"vocab-file": "/mnt/ssd-1/data/20B_tokenizer.json",
"train_data_paths": ["/mnt/ssd-1/P3/P3_combined/train_text_document"],
"valid_data_paths": ["/mnt/ssd-1/P3/P3_combined/validation_text_document"],
"test_data_paths": ["/mnt/ssd-1/P3/P3_combined/test_text_document"],
"train_data_weights": [1.0], "valid_data_weights": [1.0], "test_data_weights": [1.0],
"save": "/mnt/ssd-1/20B_P3",
"load": "/mnt/ssd-1/20B_P3",
"keep_last_n_checkpoints": 3,
"tensorboard-dir": "/mnt/ssd-1/tensorboard",
"log-dir": "/mnt/ssd-1/logs",
"wandb_team": "eleutherai",
"wandb_project": "gpt-thicc",
"wandb_group": "P3"
}
-
-
-
-
false
-
-
-
-
mmap
-
false
true
true
false
-
false
-
nccl
false
Finished
stellaathena
1d 1m 11s
-
gelu
false
1000
false
-
["global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global"]
0
false
4
-
false
true
false
-
-
true
-
false
1
-
false
1
-
# GPT-2 pretraining setup
{
# parallelism settings ( you will want to change these based on your cluster setup, ideally scheduling pipeline stages
# across the node boundaries )
"pipe-parallel-size": 4,
"model-parallel-size": 2,
# model settings
"num-layers": 44,
"hidden-size": 6144,
"num-attention-heads": 64,
"seq-length": 2048,
"max-position-embeddings": 2048,
"norm": "layernorm",
"pos-emb": "rotary",
"rotary_pct": 0.25,
"no-weight-tying": true,
"gpt_j_residual": true,
"output_layer_parallelism": "column",
# these should provide some speedup but takes a while to build, set to true if desired
"scaled-upper-triang-masked-softmax-fusion": false,
"bias-gelu-fusion": true,
# init methods
"init_method": "small_init",
"output_layer_init_method": "wang_init",
# optimizer settings
"optimizer": {
"type": "Adam",
"params": {
"lr": 9.7e-8,
"betas": [0.9, 0.95],
"eps": 1.0e-8,
}
},
"min_lr": 9.7e-8,
"override-lr-scheduler": True,
"zero_optimization": {
"stage": 1,
"allgather_partitions": True,
"allgather_bucket_size": 1260000000,
"overlap_comm": True,
"reduce_scatter": True,
"reduce_bucket_size": 1260000000,
"contiguous_gradients": True,
"cpu_offload": False
},
# batch / data settings (assuming 32 GPUs)
"train_micro_batch_size_per_gpu": 4,
"gradient_accumulation_steps": 16,
"data-impl": "mmap",
# activation checkpointing
"checkpoint-activations": true,
"checkpoint-num-layers": 1,
"partition-activations": false,
"synchronize-each-layer": true,
# regularization
"gradient_clipping": 1.0,
"weight-decay": 0.01,
"hidden-dropout": 0,
"attention-dropout": 0,
# precision settings
"fp16": {
"fp16": true,
"enabled": true,
"loss_scale": 0,
"loss_scale_window": 1000,
"initial_scale_power": 12,
"hysteresis": 2,
"min_loss_scale": 1
},
# misc. training settings
"train-iters": 200000,
"distributed-backend": "nccl",
"lr-decay-style": "constant",
"save-interval": 500,
"eval-interval": 100,
"eval-iters": 1,
# "eval_tasks": ["lambada", "piqa", "hellaswag", "winogrande", "mathqa", "pubmedqa"],
# logging
"log-interval": 2,
"steps_per_print": 2,
"wall_clock_breakdown": false,
### NEW DATA: ####
"tokenizer_type": "HFTokenizer",
"vocab-file": "/mnt/ssd-1/data/20B_tokenizer.json",
"train_data_paths": ["/mnt/ssd-1/P3/P3_combined/train_text_document"],
"valid_data_paths": ["/mnt/ssd-1/P3/P3_combined/validation_text_document"],
"test_data_paths": ["/mnt/ssd-1/P3/P3_combined/test_text_document"],
"train_data_weights": [1.0], "valid_data_weights": [1.0], "test_data_weights": [1.0],
"save": "/mnt/ssd-1/20B_P3",
"load": "/mnt/ssd-1/20B_P3",
"keep_last_n_checkpoints": 3,
"tensorboard-dir": "/mnt/ssd-1/tensorboard",
"log-dir": "/mnt/ssd-1/logs",
"wandb_team": "eleutherai",
"wandb_project": "gpt-thicc",
"wandb_group": "P3"
}
-
-
-
-
false
-
-
-
-
mmap
-
false
true
true
false
-
false
-
nccl
false
Finished
stellaathena
9h 57m 47s
-
gelu
false
1000
false
-
["global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global"]
0
false
4
-
false
true
false
-
-
true
-
false
1
-
false
1
-
# GPT-2 pretraining setup
{
# parallelism settings ( you will want to change these based on your cluster setup, ideally scheduling pipeline stages
# across the node boundaries )
"pipe-parallel-size": 4,
"model-parallel-size": 2,
# model settings
"num-layers": 44,
"hidden-size": 6144,
"num-attention-heads": 64,
"seq-length": 2048,
"max-position-embeddings": 2048,
"norm": "layernorm",
"pos-emb": "rotary",
"rotary_pct": 0.25,
"no-weight-tying": true,
"gpt_j_residual": true,
"output_layer_parallelism": "column",
# these should provide some speedup but takes a while to build, set to true if desired
"scaled-upper-triang-masked-softmax-fusion": false,
"bias-gelu-fusion": true,
# init methods
"init_method": "small_init",
"output_layer_init_method": "wang_init",
# optimizer settings
"optimizer": {
"type": "Adam",
"params": {
"lr": 9.7e-8,
"betas": [0.9, 0.95],
"eps": 1.0e-8,
}
},
"min_lr": 9.7e-8,
"override-lr-scheduler": True,
"zero_optimization": {
"stage": 1,
"allgather_partitions": True,
"allgather_bucket_size": 1260000000,
"overlap_comm": True,
"reduce_scatter": True,
"reduce_bucket_size": 1260000000,
"contiguous_gradients": True,
"cpu_offload": False
},
# batch / data settings (assuming 32 GPUs)
"train_micro_batch_size_per_gpu": 4,
"gradient_accumulation_steps": 16,
"data-impl": "mmap",
# activation checkpointing
"checkpoint-activations": true,
"checkpoint-num-layers": 1,
"partition-activations": false,
"synchronize-each-layer": true,
# regularization
"gradient_clipping": 1.0,
"weight-decay": 0.01,
"hidden-dropout": 0,
"attention-dropout": 0,
# precision settings
"fp16": {
"fp16": true,
"enabled": true,
"loss_scale": 0,
"loss_scale_window": 1000,
"initial_scale_power": 12,
"hysteresis": 2,
"min_loss_scale": 1
},
# misc. training settings
"train-iters": 200000,
"distributed-backend": "nccl",
"lr-decay-style": "constant",
"save-interval": 500,
"eval-interval": 100,
"eval-iters": 1,
# "eval_tasks": ["lambada", "piqa", "hellaswag", "winogrande", "mathqa", "pubmedqa"],
# logging
"log-interval": 2,
"steps_per_print": 2,
"wall_clock_breakdown": false,
### NEW DATA: ####
"tokenizer_type": "HFTokenizer",
"vocab-file": "/mnt/ssd-1/data/20B_tokenizer.json",
"train_data_paths": ["/mnt/ssd-1/P3/P3_combined/train_text_document"],
"valid_data_paths": ["/mnt/ssd-1/P3/P3_combined/validation_text_document"],
"test_data_paths": ["/mnt/ssd-1/P3/P3_combined/test_text_document"],
"train_data_weights": [1.0], "valid_data_weights": [1.0], "test_data_weights": [1.0],
"save": "/mnt/ssd-1/20B_P3",
"load": "/mnt/ssd-1/20B_P3",
"keep_last_n_checkpoints": 3,
"tensorboard-dir": "/mnt/ssd-1/tensorboard",
"log-dir": "/mnt/ssd-1/logs",
"wandb_team": "eleutherai",
"wandb_project": "gpt-thicc",
"wandb_group": "P3"
}
-
-
-
-
false
-
-
-
-
mmap
-
false
true
true
false
-
false
-
nccl
false
Finished
stellaathena
9h 57m 54s
-
gelu
false
1000
false
-
["global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global","global"]
0
false
4
-
false
true
false
-
-
true
-
false
1
-
false
1
-
# GPT-2 pretraining setup
{
# parallelism settings ( you will want to change these based on your cluster setup, ideally scheduling pipeline stages
# across the node boundaries )
"pipe-parallel-size": 4,
"model-parallel-size": 2,
# model settings
"num-layers": 44,
"hidden-size": 6144,
"num-attention-heads": 64,
"seq-length": 2048,
"max-position-embeddings": 2048,
"norm": "layernorm",
"pos-emb": "rotary",
"rotary_pct": 0.25,
"no-weight-tying": true,
"gpt_j_residual": true,
"output_layer_parallelism": "column",
# these should provide some speedup but takes a while to build, set to true if desired
"scaled-upper-triang-masked-softmax-fusion": false,
"bias-gelu-fusion": true,
# init methods
"init_method": "small_init",
"output_layer_init_method": "wang_init",
# optimizer settings
"optimizer": {
"type": "Adam",
"params": {
"lr": 9.7e-8,
"betas": [0.9, 0.95],
"eps": 1.0e-8,
}
},
"min_lr": 9.7e-8,
"override-lr-scheduler": True,
"zero_optimization": {
"stage": 1,
"allgather_partitions": True,
"allgather_bucket_size": 1260000000,
"overlap_comm": True,
"reduce_scatter": True,
"reduce_bucket_size": 1260000000,
"contiguous_gradients": True,
"cpu_offload": False
},
# batch / data settings (assuming 32 GPUs)
"train_micro_batch_size_per_gpu": 4,
"gradient_accumulation_steps": 16,
"data-impl": "mmap",
# activation checkpointing
"checkpoint-activations": true,
"checkpoint-num-layers": 1,
"partition-activations": false,
"synchronize-each-layer": true,
# regularization
"gradient_clipping": 1.0,
"weight-decay": 0.01,
"hidden-dropout": 0,
"attention-dropout": 0,
# precision settings
"fp16": {
"fp16": true,
"enabled": true,
"loss_scale": 0,
"loss_scale_window": 1000,
"initial_scale_power": 12,
"hysteresis": 2,
"min_loss_scale": 1
},
# misc. training settings
"train-iters": 200000,
"distributed-backend": "nccl",
"lr-decay-style": "constant",
"save-interval": 500,
"eval-interval": 100,
"eval-iters": 1,
# "eval_tasks": ["lambada", "piqa", "hellaswag", "winogrande", "mathqa", "pubmedqa"],
# logging
"log-interval": 2,
"steps_per_print": 2,
"wall_clock_breakdown": false,
### NEW DATA: ####
"tokenizer_type": "HFTokenizer",
"vocab-file": "/mnt/ssd-1/data/20B_tokenizer.json",
"train_data_paths": ["/mnt/ssd-1/P3/P3_combined/train_text_document"],
"valid_data_paths": ["/mnt/ssd-1/P3/P3_combined/validation_text_document"],
"test_data_paths": ["/mnt/ssd-1/P3/P3_combined/test_text_document"],
"train_data_weights": [1.0], "valid_data_weights": [1.0], "test_data_weights": [1.0],
"save": "/mnt/ssd-1/20B_P3",
"load": "/mnt/ssd-1/20B_P3",
"keep_last_n_checkpoints": 3,
"tensorboard-dir": "/mnt/ssd-1/tensorboard",
"log-dir": "/mnt/ssd-1/logs",
"wandb_team": "eleutherai",
"wandb_project": "gpt-thicc",
"wandb_group": "P3"
}
-
-
-
-
false
-
-
-
-
mmap
-
false
true
true
false
-
false
-
nccl
false
1-6
of 6