Upload folder using huggingface_hub
Browse files- lr2e-4_total_batch_size5120_seq_len256/args.json +1 -1
- lr2e-4_total_batch_size5120_seq_len256/dataloader_02500.pt +3 -0
- lr2e-4_total_batch_size5120_seq_len256/dataloader_04999.pt +3 -0
- lr2e-4_total_batch_size5120_seq_len256/log2.txt +0 -0
- lr2e-4_total_batch_size5120_seq_len256/model_02500.pt +3 -0
- lr2e-4_total_batch_size5120_seq_len256/model_04999.pt +3 -0
- lr2e-4_total_batch_size5120_seq_len256/optimizer_02500.pt +3 -0
- lr2e-4_total_batch_size5120_seq_len256/optimizer_04999.pt +3 -0
lr2e-4_total_batch_size5120_seq_len256/args.json
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"hellaswag": true, "attention_kind": "selective", "log_dir": "
|
|
|
|
| 1 |
+
{"hellaswag": true, "attention_kind": "selective", "log_dir": "proxy_model_sweep_2/lr2e-4_total_batch_size5120_seq_len256", "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, "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": true, "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, "n_embd": 128}
|
lr2e-4_total_batch_size5120_seq_len256/dataloader_02500.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f5f86eba64e08aa3145708d463ef689ecaee47ebe0a97035e6e0aeb59f377344
|
| 3 |
+
size 964
|
lr2e-4_total_batch_size5120_seq_len256/dataloader_04999.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:431842dc67f0d83e5eb85cf9c36fd024e295001fe78f3ba5deeae77865394feb
|
| 3 |
+
size 964
|
lr2e-4_total_batch_size5120_seq_len256/log2.txt
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
lr2e-4_total_batch_size5120_seq_len256/model_02500.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:165f2167b5ff54fb8d7130360578d22681f65fd69f0adb84d44f5d57a98a2d7f
|
| 3 |
+
size 38587970
|
lr2e-4_total_batch_size5120_seq_len256/model_04999.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:38abd7e541ba3071c2f74093b3d81336901abd7362822f844be7642c7b399630
|
| 3 |
+
size 38587970
|
lr2e-4_total_batch_size5120_seq_len256/optimizer_02500.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:62cef25c61230a318310a93cceab76b6f8d1bdd5efc61adb8b36f779c5c6eb61
|
| 3 |
+
size 70895430
|
lr2e-4_total_batch_size5120_seq_len256/optimizer_04999.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:08b782a124fc1ce8b3671cec420a267708a8b34cb5fbd352f84022fe7bf6b498
|
| 3 |
+
size 70895430
|