Hugging Face's logo

andrew-healey
/
context-compression

Model card Files
xet
Community
andrew-healey commited on
Commit
7a14b7b
·
verified ·
1 Parent(s): 80fd823

Upload folder using huggingface_hub

Browse files
Files changed (14) hide show
  1. lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/args.json +1 -0
  2. lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/dataloader_02500.pt +3 -0
  3. lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/dataloader_05000.pt +3 -0
  4. lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/dataloader_07500.pt +3 -0
  5. lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/dataloader_09999.pt +3 -0
  6. lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/log2.txt +0 -0
  7. lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/model_02500.pt +3 -0
  8. lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/model_05000.pt +3 -0
  9. lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/model_07500.pt +3 -0
  10. lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/model_09999.pt +3 -0
  11. lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/optimizer_02500.pt +3 -0
  12. lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/optimizer_05000.pt +3 -0
  13. lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/optimizer_07500.pt +3 -0
  14. lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/optimizer_09999.pt +3 -0
lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/args.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"hellaswag": true, "attention_kind": "selective", "log_dir": "proxy_model_sweep_2/lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1", "resume_checkpoint": null, "resume_optimizer": false, "add_a_head": false, "add_head_to_start": true, "new_head_init": "normal", "n_heads": 2, "protect_bos_token": true, "prevent_from_masking_myself": true, "max_steps": 10000, "warmup_steps": 1, "group": "proxy_model_sweep_2", "use_wandb": true, "kill_self_after_run": false, "random_seed": 1337, "memory_penalty_epsilon": 0.1, "selection_head_linear_combo": "none", "selection_head_linear_combo_scale": 1.0, "protection_kind": "none", "leaky_relu_alpha": null, "leaky_relu_bias": null, "use_compile": true, "use_mini_model": false, "upload_to_hf": true, "seq_len": 256, "batch_size": 20, "total_batch_size": 5120, "protection_head_scaling_factor": 1.0, "protection_head_bias": 0.0, "n_sliced_masks": null, "n_latent_masks": null, "mask_layernorm": false, "residual_attention_masks": false, "compute_base_shapes": false, "base_shapes_savefile": null, "mup": true, "disable_selection": false, "mup_enable_coord_check_logging": false, "max_lr": 0.000175, "decay_lr": true, "readout_zero_init": false, "query_zero_init": false, "l1_loss": false, "debugpy": false, "n_embd": 128}
lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/dataloader_02500.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5f86eba64e08aa3145708d463ef689ecaee47ebe0a97035e6e0aeb59f377344
3
+ size 964
lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/dataloader_05000.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a669cff46eb6a9ac053ccb64aba9cabfe25f09b9de8be0fa7cbcf53d547468bd
3
+ size 964
lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/dataloader_07500.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d5ccc252bab74018a361c32a7ec3feffb7cfbeb3a9386972640ad89c08895266
3
+ size 964
lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/dataloader_09999.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2b95c984402cc9cb954a35bd3e3a1d7e71c550bc1628b6c4dca17aedfa872fb2
3
+ size 964
lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/log2.txt ADDED
The diff for this file is too large to render. See raw diff
 
lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/model_02500.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4db6fa30dbbb8c2648218f4b72aebf05b842e6e86a099e6e9e8f189ef160f389
3
+ size 38587970
lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/model_05000.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cfad97a57d3a17da622b5256ed0437f0605ce440ce29635de46737dcfc8952bb
3
+ size 38587970
lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/model_07500.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0cef1d9fccd384ca8579ce766ccac8cffc1a000a0bc291ca4d9ce695cc5c7a5d
3
+ size 38587970
lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/model_09999.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f74f2bdd1e295e3b28349a3ab741a07f769f2311c4c83bf114363b53e65ebc0
3
+ size 38587970
lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/optimizer_02500.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a2bb800e7fd627375a06373c84133e8a8066e722b915d9e635e127ca410306fc
3
+ size 70895430
lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/optimizer_05000.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ed9f39127bb1239485cb3fc993e33c95452d7cdcbd26843b3e8db62cba57babd
3
+ size 70895430
lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/optimizer_07500.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d72628188484e0c8f42d104d1a07f18614432e78855b5be822429865e4b84a5d
3
+ size 70895430
lr1.75e-4_total_batch_size5120_seq_len256_decay_lrtrue_attention_kindselective_warmup_steps1/optimizer_09999.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:01ace8edd6f82273317c7c55731bd5c0d724b215df8e35c3093d21e529d6d576
3
+ size 70895430