Hugging Face's logo

andrew-healey
/
context-compression

Model card Files
xet
Community
andrew-healey commited on
Commit
3c90a38
·
verified ·
1 Parent(s): af53c2b

Upload folder using huggingface_hub

Browse files
Files changed (8) hide show
  1. lr1.5e-4_total_batch_size5120_n_heads2_seed1339/args.json +1 -0
  2. lr1.5e-4_total_batch_size5120_n_heads2_seed1339/dataloader_02500.pt +3 -0
  3. lr1.5e-4_total_batch_size5120_n_heads2_seed1339/dataloader_04999.pt +3 -0
  4. lr1.5e-4_total_batch_size5120_n_heads2_seed1339/log2.txt +0 -0
  5. lr1.5e-4_total_batch_size5120_n_heads2_seed1339/model_02500.pt +3 -0
  6. lr1.5e-4_total_batch_size5120_n_heads2_seed1339/model_04999.pt +3 -0
  7. lr1.5e-4_total_batch_size5120_n_heads2_seed1339/optimizer_02500.pt +3 -0
  8. lr1.5e-4_total_batch_size5120_n_heads2_seed1339/optimizer_04999.pt +3 -0
lr1.5e-4_total_batch_size5120_n_heads2_seed1339/args.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"hellaswag": true, "attention_kind": "selective", "log_dir": "proxy_model_sweep_3/lr1.5e-4_total_batch_size5120_n_heads2_seed1339", "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": 5000, "warmup_steps": 200, "group": "proxy_model_sweep_3", "use_wandb": true, "kill_self_after_run": false, "random_seed": 1339, "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": 0.00015, "decay_lr": true, "readout_zero_init": false, "query_zero_init": false, "l1_loss": false, "debugpy": false, "n_embd": 128}
lr1.5e-4_total_batch_size5120_n_heads2_seed1339/dataloader_02500.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5f86eba64e08aa3145708d463ef689ecaee47ebe0a97035e6e0aeb59f377344
3
+ size 964
lr1.5e-4_total_batch_size5120_n_heads2_seed1339/dataloader_04999.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:431842dc67f0d83e5eb85cf9c36fd024e295001fe78f3ba5deeae77865394feb
3
+ size 964
lr1.5e-4_total_batch_size5120_n_heads2_seed1339/log2.txt ADDED
The diff for this file is too large to render. See raw diff
 
lr1.5e-4_total_batch_size5120_n_heads2_seed1339/model_02500.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0fbddfb3722f93f3b3b102bcc897ab1aefda47154caa50349e8375d254e399db
3
+ size 38587970
lr1.5e-4_total_batch_size5120_n_heads2_seed1339/model_04999.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fef61c08629376302cbbaba56e649b4702567b9d4d9becc3d6dec170220e2983
3
+ size 38587970
lr1.5e-4_total_batch_size5120_n_heads2_seed1339/optimizer_02500.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fa16e56609d946631fefcab2fdd53a7699d46e01ec3f640d3e3da392bee3261b
3
+ size 70895430
lr1.5e-4_total_batch_size5120_n_heads2_seed1339/optimizer_04999.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ca452417ff0efcf87a8cb6b1e8b0d5f2d92b49ff3d3fd73ceda16eb22de2441f
3
+ size 70895430