Hugging Face's logo

andrew-healey
/
context-compression

Model card Files
xet
 Community
andrew-healey commited on
Commit
b2a56f1
·
verified ·
1 Parent(s): 86d3c04

Upload folder using huggingface_hub

Browse files
Files changed (8) hide show
  1. lr2e-4_total_batch_size10240_n_heads2_seed1338/args.json +1 -0
  2. lr2e-4_total_batch_size10240_n_heads2_seed1338/dataloader_02500.pt +3 -0
  3. lr2e-4_total_batch_size10240_n_heads2_seed1338/dataloader_04999.pt +3 -0
  4. lr2e-4_total_batch_size10240_n_heads2_seed1338/log2.txt +0 -0
  5. lr2e-4_total_batch_size10240_n_heads2_seed1338/model_02500.pt +3 -0
  6. lr2e-4_total_batch_size10240_n_heads2_seed1338/model_04999.pt +3 -0
  7. lr2e-4_total_batch_size10240_n_heads2_seed1338/optimizer_02500.pt +3 -0
  8. lr2e-4_total_batch_size10240_n_heads2_seed1338/optimizer_04999.pt +3 -0
lr2e-4_total_batch_size10240_n_heads2_seed1338/args.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"hellaswag": true, "attention_kind": "selective", "log_dir": "proxy_model_sweep_3/lr2e-4_total_batch_size10240_n_heads2_seed1338", "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": 1338, "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": 10240, "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": 0.0002, "decay_lr": true, "readout_zero_init": false, "query_zero_init": false, "l1_loss": false, "debugpy": false, "key": "2e-4_10240_2_1338", "n_embd": 128}
lr2e-4_total_batch_size10240_n_heads2_seed1338/dataloader_02500.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8b2ea67f78ff5a7970d0db044ff7ee527b3dc065f295fd30f588df4b44b568d0
3
+ size 964
lr2e-4_total_batch_size10240_n_heads2_seed1338/dataloader_04999.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3fbc8f5b7d2476463a241dae384ac290c173f75d26d125634e27c49e7ea5399c
3
+ size 964
lr2e-4_total_batch_size10240_n_heads2_seed1338/log2.txt ADDED
The diff for this file is too large to render. See raw diff
 
lr2e-4_total_batch_size10240_n_heads2_seed1338/model_02500.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e90d43e58f30d0155bb4c6027de2696e0d56552f6827d159b8d53cb2ad1c3d3b
3
+ size 38587970
lr2e-4_total_batch_size10240_n_heads2_seed1338/model_04999.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2504ddc5cb16445563d030022c08f1c56a5c1b379344009e02929edaa3129a41
3
+ size 38587970
lr2e-4_total_batch_size10240_n_heads2_seed1338/optimizer_02500.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bfc50cc11f5ff0c8f1cc85cec327bda0ca59f95b9a883ebdb7d26a0f58ff1f5e
3
+ size 70895430
lr2e-4_total_batch_size10240_n_heads2_seed1338/optimizer_04999.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fb89ec4af19bdc21e91a31eed5d9ac3a09a04e45c91c12afba5d7192868f5655
3
+ size 70895430