diff --git a/README.md b/README.md index 081e039..1833b1c 100644 --- a/README.md +++ b/README.md @@ -103,7 +103,7 @@ gsutil -m rsync -r data/raw_datasets/self_gen gs://ctx-to-lora/data/raw_datasets mkdir -p data/raw_datasets/self_gen gsutil -m rsync -r gs://ctx-to-lora/data/raw_datasets/self_gen data/raw_datasets/self_gen -# self-gen eval +# download only self-gen eval mkdir -p data/raw_datasets/self_gen/google/gemma-2-2b-it_temp_0.0_closed_qa_prob_0.0/fw_qa_v2/min_0_to_2000/train/ gsutil -m cp -r gs://ctx-to-lora/data/raw_datasets/self_gen/google/gemma-2-2b-it_temp_0.0_closed_qa_prob_0.0/fw_qa_v2/min_0_to_2000/train/*level_0_val*.parquet data/raw_datasets/self_gen/google/gemma-2-2b-it_temp_0.0_closed_qa_prob_0.0/fw_qa_v2/min_0_to_2000/train/ ``` diff --git a/data/generate_ctx_magic_number.py b/data/generate_ctx_magic_number.py index 23b05d0..482b45e 100644 --- a/data/generate_ctx_magic_number.py +++ b/data/generate_ctx_magic_number.py @@ -169,6 +169,7 @@ def main(): (1024 * i, 1024 * (i + 1)) for i in range(1, 16) ] tok_bins += [(2**14 + 2**12 * (i), 2**14 + 2**12 * (i + 1)) for i in range(4)] + tok_bins += [(2**15 + 2**13 * (i), 2**15 + 2**13 * (i + 1)) for i in range(12)] if args.only_first_n_bins is not None: tok_bins = tok_bins[: args.only_first_n_bins] diff --git a/scripts/toy_exp.sh b/scripts/toy_exp.sh index 2cea0e2..6f1b7cb 100644 --- a/scripts/toy_exp.sh +++ b/scripts/toy_exp.sh @@ -2,10 +2,10 @@ uv run data/generate_ctx_magic_num.py # train -WANDB_PROJECT=ctx-magic-num srun --partition=aiscilow --gpus=1 --unbuffered uv run train.py configs/toy_exp/ctx_magic_number_32_256.yaml --model_name_or_path=google/gemma-2-2b-it --num_train_epochs=1 --per_device_train_batch_size=-1 --gradient_accumulation_steps=32 --per_device_eval_batch_size=16 --exp_setup=hyper_lora --aggregator_type=perceiver --target_modules=down_proj --num_blocks=8 --num_self_attn_per_block=0 --num_pre_head_layers=1 --lora_r=8 --eval_steps=100 --save_steps=1000 --learning_rate=4e-5 --lora_dropout=0.0 --neftune_noise_alpha=0 --per_rank_gen=True --per_layer_processing=True --gen_lora_l1_reg_coef=0.1 --use_sequence_packing=True --max_packed_inp_len=2048 --max_packed_ctx_len=2048 --dataloader_num_workers=0 --dataloader_prefetch_factor=None --eval_on_start=False --ctx_encoder_type=early_exit --n_latent_queries=208 --use_kl_loss=False --eval_on_start=True --lora_r=8 --max_ctx_chunk_len=-1 --max_val_samples_per_ds=100 --seed=1 -WANDB_PROJECT=ctx-magic-num srun --partition=aiscilow --gpus=1 --unbuffered uv run train.py configs/toy_exp/ctx_magic_number_32_256.yaml --model_name_or_path=google/gemma-2-2b-it --num_train_epochs=1 --per_device_train_batch_size=-1 --gradient_accumulation_steps=32 --per_device_eval_batch_size=16 --exp_setup=hyper_lora --aggregator_type=perceiver --target_modules=down_proj --num_blocks=8 --num_self_attn_per_block=0 --num_pre_head_layers=1 --lora_r=8 --eval_steps=100 --save_steps=1000 --learning_rate=4e-5 --lora_dropout=0.0 --neftune_noise_alpha=0 --per_rank_gen=True --per_layer_processing=True --gen_lora_l1_reg_coef=0.1 --use_sequence_packing=True --max_packed_inp_len=2048 --max_packed_ctx_len=2048 --dataloader_num_workers=0 --dataloader_prefetch_factor=None --eval_on_start=False --ctx_encoder_type=early_exit --n_latent_queries=208 --use_kl_loss=False --eval_on_start=True --lora_r=8 --max_ctx_chunk_len=512 --min_ctx_chunk_len=25 --num_chunk_probs='{"1":"0.5", "2":"0.5"}' --max_val_samples_per_ds=100 --seed=1 -WANDB_PROJECT=ctx-magic-num srun --partition=aiscilow --gpus=1 --unbuffered uv run train.py configs/toy_exp/ctx_magic_number_32_256.yaml --model_name_or_path=google/gemma-2-2b-it --num_train_epochs=1 --per_device_train_batch_size=-1 --gradient_accumulation_steps=32 --per_device_eval_batch_size=16 --exp_setup=hyper_lora --aggregator_type=perceiver --target_modules=down_proj --num_blocks=8 --num_self_attn_per_block=0 --num_pre_head_layers=1 --lora_r=8 --eval_steps=100 --save_steps=1000 --learning_rate=4e-5 --lora_dropout=0.0 --neftune_noise_alpha=0 --per_rank_gen=True --per_layer_processing=True --gen_lora_l1_reg_coef=0.1 --use_sequence_packing=True --max_packed_inp_len=2048 --max_packed_ctx_len=2048 --dataloader_num_workers=0 --dataloader_prefetch_factor=None --eval_on_start=False --ctx_encoder_type=early_exit --n_latent_queries=208 --use_kl_loss=False --eval_on_start=True --lora_r=8 --max_ctx_chunk_len=512 --min_ctx_chunk_len=25 --num_chunk_probs='{"1":"0.5", "2":"0.25", "3":"0.125", "4":"0.125"}' --max_val_samples_per_ds=100 --seed=1 -WANDB_PROJECT=ctx-magic-num run uv run train.py configs/toy_exp/ctx_magic_number_32_256.yaml --model_name_or_path=google/gemma-2-2b-it --num_train_epochs=1 --per_device_train_batch_size=-1 --gradient_accumulation_steps=32 --per_device_eval_batch_size=16 --exp_setup=hyper_lora --aggregator_type=perceiver --target_modules=down_proj --num_blocks=8 --num_self_attn_per_block=0 --num_pre_head_layers=1 --lora_r=8 --eval_steps=100 --save_steps=1000 --learning_rate=4e-5 --lora_dropout=0.0 --neftune_noise_alpha=0 --per_rank_gen=True --per_layer_processing=True --gen_lora_l1_reg_coef=0.1 --use_sequence_packing=True --max_packed_inp_len=2048 --max_packed_ctx_len=2048 --dataloader_num_workers=0 --dataloader_prefetch_factor=None --eval_on_start=False --ctx_encoder_type=early_exit --n_latent_queries=208 --use_kl_loss=False --eval_on_start=True --lora_r=8 --max_ctx_chunk_len=512 --min_ctx_chunk_len=25 --num_chunk_probs='{"1":"0.5", "2":"0.125", "3":"0.0625", "4":"0.0625", "5":"0.0625", "6":"0.0625", "7":"0.0625", "8":"0.0625"}' --max_val_samples_per_ds=100 --seed=1 +# WANDB_PROJECT=ctx-magic-num srun --partition=aiscilow --gpus=1 --unbuffered uv run train.py configs/toy_exp/ctx_magic_number_32_256.yaml --model_name_or_path=google/gemma-2-2b-it --num_train_epochs=1 --per_device_train_batch_size=-1 --gradient_accumulation_steps=32 --per_device_eval_batch_size=16 --exp_setup=hyper_lora --aggregator_type=perceiver --target_modules=down_proj --num_blocks=8 --num_self_attn_per_block=0 --num_pre_head_layers=1 --lora_r=8 --eval_steps=100 --save_steps=1000 --learning_rate=4e-5 --lora_dropout=0.0 --neftune_noise_alpha=0 --per_rank_gen=True --per_layer_processing=True --gen_lora_l1_reg_coef=0.1 --use_sequence_packing=True --max_packed_inp_len=2048 --max_packed_ctx_len=2048 --dataloader_num_workers=0 --dataloader_prefetch_factor=None --eval_on_start=False --ctx_encoder_type=early_exit --n_latent_queries=208 --use_kl_loss=False --eval_on_start=True --lora_r=8 --max_ctx_chunk_len=-1 --max_val_samples_per_ds=100 --seed=1 +# WANDB_PROJECT=ctx-magic-num srun --partition=aiscilow --gpus=1 --unbuffered uv run train.py configs/toy_exp/ctx_magic_number_32_256.yaml --model_name_or_path=google/gemma-2-2b-it --num_train_epochs=1 --per_device_train_batch_size=-1 --gradient_accumulation_steps=32 --per_device_eval_batch_size=16 --exp_setup=hyper_lora --aggregator_type=perceiver --target_modules=down_proj --num_blocks=8 --num_self_attn_per_block=0 --num_pre_head_layers=1 --lora_r=8 --eval_steps=100 --save_steps=1000 --learning_rate=4e-5 --lora_dropout=0.0 --neftune_noise_alpha=0 --per_rank_gen=True --per_layer_processing=True --gen_lora_l1_reg_coef=0.1 --use_sequence_packing=True --max_packed_inp_len=2048 --max_packed_ctx_len=2048 --dataloader_num_workers=0 --dataloader_prefetch_factor=None --eval_on_start=False --ctx_encoder_type=early_exit --n_latent_queries=208 --use_kl_loss=False --eval_on_start=True --lora_r=8 --max_ctx_chunk_len=512 --min_ctx_chunk_len=25 --num_chunk_probs='{"1":"0.5", "2":"0.5"}' --max_val_samples_per_ds=100 --seed=1 +# WANDB_PROJECT=ctx-magic-num srun --partition=aiscilow --gpus=1 --unbuffered uv run train.py configs/toy_exp/ctx_magic_number_32_256.yaml --model_name_or_path=google/gemma-2-2b-it --num_train_epochs=1 --per_device_train_batch_size=-1 --gradient_accumulation_steps=32 --per_device_eval_batch_size=16 --exp_setup=hyper_lora --aggregator_type=perceiver --target_modules=down_proj --num_blocks=8 --num_self_attn_per_block=0 --num_pre_head_layers=1 --lora_r=8 --eval_steps=100 --save_steps=1000 --learning_rate=4e-5 --lora_dropout=0.0 --neftune_noise_alpha=0 --per_rank_gen=True --per_layer_processing=True --gen_lora_l1_reg_coef=0.1 --use_sequence_packing=True --max_packed_inp_len=2048 --max_packed_ctx_len=2048 --dataloader_num_workers=0 --dataloader_prefetch_factor=None --eval_on_start=False --ctx_encoder_type=early_exit --n_latent_queries=208 --use_kl_loss=False --eval_on_start=True --lora_r=8 --max_ctx_chunk_len=512 --min_ctx_chunk_len=25 --num_chunk_probs='{"1":"0.5", "2":"0.25", "3":"0.125", "4":"0.125"}' --max_val_samples_per_ds=100 --seed=1 +WANDB_PROJECT=ctx-magic-num run uv run train.py configs/toy_exp/ctx_magic_number_32_256.yaml --model_name_or_path=google/gemma-2-2b-it --num_train_epochs=1 --per_device_train_batch_size=-1 --gradient_accumulation_steps=16 --per_device_eval_batch_size=16 --exp_setup=hyper_lora --aggregator_type=perceiver --target_modules=down_proj --num_blocks=8 --num_self_attn_per_block=0 --num_pre_head_layers=1 --lora_r=8 --eval_steps=100 --logging_steps=10 --save_steps=1000 --learning_rate=4e-5 --lora_dropout=0.0 --neftune_noise_alpha=0 --per_rank_gen=True --per_layer_processing=True --gen_lora_l1_reg_coef=1 --use_sequence_packing=True --max_packed_inp_len=4096 --max_packed_ctx_len=4096 --dataloader_num_workers=0 --dataloader_prefetch_factor=None --eval_on_start=False --ctx_encoder_type=early_exit --n_latent_queries=208 --use_kl_loss=False --eval_on_start=True --lora_r=8 --max_ctx_chunk_len=512 --min_ctx_chunk_len=25 --num_chunk_probs='{"1":"0.5", "2":"0.125", "3":"0.0625", "4":"0.0625", "5":"0.0625", "6":"0.0625", "7":"0.0625", "8":"0.0625"}' --max_val_samples_per_ds=100 --seed=1 # eval -WANDB_MODE=disabled srun --partition=aiscilow --gpus=1 --unbuffered uv run run_eval.py --checkpoint_path CHECKPOINT_PATH --datasets ctx_magic_number_32_1024 ctx_magic_number_1024_2048 ctx_magic_number_2048_3072 ctx_magic_number_3072_4096 ctx_magic_number_4096_5120 ctx_magic_number_5120_6144 ctx_magic_number_6144_7168 ctx_magic_number_7168_8192 ctx_magic_number_8192_9216 ctx_magic_number_9216_10240 ctx_magic_number_10240_11264 ctx_magic_number_11264_12288 ctx_magic_number_12288_13312 ctx_magic_number_13312_14336 ctx_magic_number_14336_15360 ctx_magic_number_15360_16384 ctx_magic_number_16384_20480 ctx_magic_number_20480_24576 ctx_magic_number_24576_28672 ctx_magic_number_28672_32768 --max_ctx_chunk_len=1024 --split test +WANDB_MODE=disabled srun --partition=aiscilow --gpus=1 --unbuffered uv run run_eval.py --checkpoint_path CHECKPOINT_PATH --datasets ctx_magic_number_32_1024 ctx_magic_number_1024_2048 ctx_magic_number_2048_3072 ctx_magic_number_3072_4096 ctx_magic_number_4096_5120 ctx_magic_number_5120_6144 ctx_magic_number_6144_7168 ctx_magic_number_7168_8192 ctx_magic_number_8192_9216 ctx_magic_number_9216_10240 ctx_magic_number_10240_11264 ctx_magic_number_11264_12288 ctx_magic_number_12288_13312 ctx_magic_number_13312_14336 ctx_magic_number_14336_15360 ctx_magic_number_15360_16384 ctx_magic_number_16384_20480 ctx_magic_number_20480_24576 ctx_magic_number_24576_28672 ctx_magic_number_28672_32768 ctx_magic_number_32768_40960 ctx_magic_number_40960_49152 ctx_magic_number_49152_57344 ctx_magic_number_57344_65536 ctx_magic_number_65536_73728 ctx_magic_number_73728_81920 ctx_magic_number_81920_90112 ctx_magic_number_90112_98304 ctx_magic_number_98304_106496 ctx_magic_number_106496_114688 ctx_magic_number_114688_122880 ctx_magic_number_122880_131072 --max_ctx_chunk_len=1024 --split test --eval_batch_size_gen=4 diff --git a/src/ctx_to_lora/data/definitions.py b/src/ctx_to_lora/data/definitions.py index 6afaf15..d4faa7b 100644 --- a/src/ctx_to_lora/data/definitions.py +++ b/src/ctx_to_lora/data/definitions.py @@ -699,6 +699,7 @@ tok_bins += [(32, 128), (128, 256), (256, 512), (512, 1024), (32, 1024)] + [ (1024 * i, 1024 * (i + 1)) for i in range(1, 16) ] tok_bins += [(2**14 + 2**12 * (i), 2**14 + 2**12 * (i + 1)) for i in range(4)] +tok_bins += [(2**15 + 2**13 * (i), 2**15 + 2**13 * (i + 1)) for i in range(12)] for toy_ds_name in ["ctx_numbers", "ctx_kv", "ctx_magic_number"]: for tok_bin in tok_bins: DS_KWARGS[f"{toy_ds_name}_{tok_bin[0]}_{tok_bin[1]}"] = dict(