Training in progress, step 500, checkpoint
Browse files- last-checkpoint/adapter_model.safetensors +1 -1
- last-checkpoint/global_step500/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt +3 -0
- last-checkpoint/global_step500/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt +3 -0
- last-checkpoint/global_step500/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt +3 -0
- last-checkpoint/global_step500/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt +3 -0
- last-checkpoint/global_step500/zero_pp_rank_0_mp_rank_00_model_states.pt +3 -0
- last-checkpoint/global_step500/zero_pp_rank_1_mp_rank_00_model_states.pt +3 -0
- last-checkpoint/global_step500/zero_pp_rank_2_mp_rank_00_model_states.pt +3 -0
- last-checkpoint/global_step500/zero_pp_rank_3_mp_rank_00_model_states.pt +3 -0
- last-checkpoint/latest +1 -1
- last-checkpoint/rng_state_0.pth +1 -1
- last-checkpoint/rng_state_1.pth +1 -1
- last-checkpoint/rng_state_2.pth +1 -1
- last-checkpoint/rng_state_3.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +727 -3
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13648688
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:14728c0dda1002b24225273f56807d0922c03feb73afbc3c9a2b84f92ba99853
|
| 3 |
size 13648688
|
last-checkpoint/global_step500/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:945c95b87cfa9a4247bbc67dec060d40f98d220efeb8d7bfea909d9f47efb568
|
| 3 |
+
size 20450800
|
last-checkpoint/global_step500/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dae0e38e3cbaf2a462484d1476859c6d7867323c4af89510eb89d5c9d08aed63
|
| 3 |
+
size 20450800
|
last-checkpoint/global_step500/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b5e026b69b255260812d62d318537ba4640d841fca11b458b154d3a31b03c0af
|
| 3 |
+
size 20450800
|
last-checkpoint/global_step500/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6ec4962a539f87f3a90b500e4e367939ad8029ef0f09c092166c135e5280ad6d
|
| 3 |
+
size 20450800
|
last-checkpoint/global_step500/zero_pp_rank_0_mp_rank_00_model_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:069c1f3451602986b71542c8dfb178a4f03afe6c9b0c4ee35743ce4c002eee59
|
| 3 |
+
size 152238
|
last-checkpoint/global_step500/zero_pp_rank_1_mp_rank_00_model_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cfa039f9de21be98beb1eed1119437f237d6f512e1b762fe0af46a5e4792c3a3
|
| 3 |
+
size 152238
|
last-checkpoint/global_step500/zero_pp_rank_2_mp_rank_00_model_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fe9a3416e998f9b31c1d15894e49a2114183977f218cd3bfba00c3afc6923016
|
| 3 |
+
size 152238
|
last-checkpoint/global_step500/zero_pp_rank_3_mp_rank_00_model_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:29695eef268967bbc32ca3feaf63f2a35ba3ac928a655cbd7d49855450e24ca3
|
| 3 |
+
size 152238
|
last-checkpoint/latest
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
|
|
|
|
| 1 |
+
global_step500
|
last-checkpoint/rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14960
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3f1c62d3591874d9423058ca8b4ee1529ba22e4a464cd7a05c124170a54d10f7
|
| 3 |
size 14960
|
last-checkpoint/rng_state_1.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14960
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:769adabc2930b4c5ffaf06e4c14dcf6bc475cf3777b500b26f6eee3ded6fabd2
|
| 3 |
size 14960
|
last-checkpoint/rng_state_2.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14960
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1bbf4e280aa45bd7ecd510773cee429432a2cf8a426b57fcb1765c6e16762915
|
| 3 |
size 14960
|
last-checkpoint/rng_state_3.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14960
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f776eff2202474e4c474585023a50e014369bc49fa31bbd19616214366af09f8
|
| 3 |
size 14960
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6b3f9425ed43b96686ab768302a7b350060720823822160ec86dcae8519a36a4
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 40,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2939,6 +2939,730 @@
|
|
| 2939 |
"eval_samples_per_second": 2.201,
|
| 2940 |
"eval_steps_per_second": 0.176,
|
| 2941 |
"step": 400
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2942 |
}
|
| 2943 |
],
|
| 2944 |
"logging_steps": 1,
|
|
@@ -2958,7 +3682,7 @@
|
|
| 2958 |
"attributes": {}
|
| 2959 |
}
|
| 2960 |
},
|
| 2961 |
-
"total_flos":
|
| 2962 |
"train_batch_size": 4,
|
| 2963 |
"trial_name": null,
|
| 2964 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.6148170919151552,
|
| 5 |
"eval_steps": 40,
|
| 6 |
+
"global_step": 500,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2939 |
"eval_samples_per_second": 2.201,
|
| 2940 |
"eval_steps_per_second": 0.176,
|
| 2941 |
"step": 400
|
| 2942 |
+
},
|
| 2943 |
+
{
|
| 2944 |
+
"epoch": 0.4930833077159545,
|
| 2945 |
+
"grad_norm": 0.4925378998198651,
|
| 2946 |
+
"learning_rate": 1.1985138830231638e-05,
|
| 2947 |
+
"loss": 0.2452,
|
| 2948 |
+
"step": 401
|
| 2949 |
+
},
|
| 2950 |
+
{
|
| 2951 |
+
"epoch": 0.4943129418997848,
|
| 2952 |
+
"grad_norm": 0.43350371819373273,
|
| 2953 |
+
"learning_rate": 1.1942999301461694e-05,
|
| 2954 |
+
"loss": 0.2145,
|
| 2955 |
+
"step": 402
|
| 2956 |
+
},
|
| 2957 |
+
{
|
| 2958 |
+
"epoch": 0.4955425760836151,
|
| 2959 |
+
"grad_norm": 0.4788710711794286,
|
| 2960 |
+
"learning_rate": 1.1900823885707216e-05,
|
| 2961 |
+
"loss": 0.2355,
|
| 2962 |
+
"step": 403
|
| 2963 |
+
},
|
| 2964 |
+
{
|
| 2965 |
+
"epoch": 0.49677221026744545,
|
| 2966 |
+
"grad_norm": 0.48489068989863693,
|
| 2967 |
+
"learning_rate": 1.1858613361943518e-05,
|
| 2968 |
+
"loss": 0.1636,
|
| 2969 |
+
"step": 404
|
| 2970 |
+
},
|
| 2971 |
+
{
|
| 2972 |
+
"epoch": 0.49800184445127577,
|
| 2973 |
+
"grad_norm": 0.38739819249024443,
|
| 2974 |
+
"learning_rate": 1.1816368509794365e-05,
|
| 2975 |
+
"loss": 0.1768,
|
| 2976 |
+
"step": 405
|
| 2977 |
+
},
|
| 2978 |
+
{
|
| 2979 |
+
"epoch": 0.49923147863510603,
|
| 2980 |
+
"grad_norm": 0.6455001321481116,
|
| 2981 |
+
"learning_rate": 1.177409010951755e-05,
|
| 2982 |
+
"loss": 0.2625,
|
| 2983 |
+
"step": 406
|
| 2984 |
+
},
|
| 2985 |
+
{
|
| 2986 |
+
"epoch": 0.5004611128189363,
|
| 2987 |
+
"grad_norm": 0.6173490831722124,
|
| 2988 |
+
"learning_rate": 1.1731778941990497e-05,
|
| 2989 |
+
"loss": 0.1915,
|
| 2990 |
+
"step": 407
|
| 2991 |
+
},
|
| 2992 |
+
{
|
| 2993 |
+
"epoch": 0.5016907470027667,
|
| 2994 |
+
"grad_norm": 0.4814901038602925,
|
| 2995 |
+
"learning_rate": 1.1689435788695844e-05,
|
| 2996 |
+
"loss": 0.1646,
|
| 2997 |
+
"step": 408
|
| 2998 |
+
},
|
| 2999 |
+
{
|
| 3000 |
+
"epoch": 0.502920381186597,
|
| 3001 |
+
"grad_norm": 0.4617832240658794,
|
| 3002 |
+
"learning_rate": 1.1647061431707e-05,
|
| 3003 |
+
"loss": 0.1671,
|
| 3004 |
+
"step": 409
|
| 3005 |
+
},
|
| 3006 |
+
{
|
| 3007 |
+
"epoch": 0.5041500153704273,
|
| 3008 |
+
"grad_norm": 0.31842872427334545,
|
| 3009 |
+
"learning_rate": 1.1604656653673707e-05,
|
| 3010 |
+
"loss": 0.1587,
|
| 3011 |
+
"step": 410
|
| 3012 |
+
},
|
| 3013 |
+
{
|
| 3014 |
+
"epoch": 0.5053796495542576,
|
| 3015 |
+
"grad_norm": 0.4310369489277566,
|
| 3016 |
+
"learning_rate": 1.156222223780757e-05,
|
| 3017 |
+
"loss": 0.2058,
|
| 3018 |
+
"step": 411
|
| 3019 |
+
},
|
| 3020 |
+
{
|
| 3021 |
+
"epoch": 0.5066092837380879,
|
| 3022 |
+
"grad_norm": 0.3434568907080096,
|
| 3023 |
+
"learning_rate": 1.1519758967867608e-05,
|
| 3024 |
+
"loss": 0.1748,
|
| 3025 |
+
"step": 412
|
| 3026 |
+
},
|
| 3027 |
+
{
|
| 3028 |
+
"epoch": 0.5078389179219183,
|
| 3029 |
+
"grad_norm": 0.3212603412265404,
|
| 3030 |
+
"learning_rate": 1.1477267628145777e-05,
|
| 3031 |
+
"loss": 0.1998,
|
| 3032 |
+
"step": 413
|
| 3033 |
+
},
|
| 3034 |
+
{
|
| 3035 |
+
"epoch": 0.5090685521057485,
|
| 3036 |
+
"grad_norm": 0.3727679962149267,
|
| 3037 |
+
"learning_rate": 1.1434749003452467e-05,
|
| 3038 |
+
"loss": 0.1962,
|
| 3039 |
+
"step": 414
|
| 3040 |
+
},
|
| 3041 |
+
{
|
| 3042 |
+
"epoch": 0.5102981862895789,
|
| 3043 |
+
"grad_norm": 0.5738583382215159,
|
| 3044 |
+
"learning_rate": 1.1392203879102027e-05,
|
| 3045 |
+
"loss": 0.1792,
|
| 3046 |
+
"step": 415
|
| 3047 |
+
},
|
| 3048 |
+
{
|
| 3049 |
+
"epoch": 0.5115278204734092,
|
| 3050 |
+
"grad_norm": 0.3242461291494194,
|
| 3051 |
+
"learning_rate": 1.1349633040898246e-05,
|
| 3052 |
+
"loss": 0.1407,
|
| 3053 |
+
"step": 416
|
| 3054 |
+
},
|
| 3055 |
+
{
|
| 3056 |
+
"epoch": 0.5127574546572394,
|
| 3057 |
+
"grad_norm": 0.39000456676333345,
|
| 3058 |
+
"learning_rate": 1.1307037275119854e-05,
|
| 3059 |
+
"loss": 0.1644,
|
| 3060 |
+
"step": 417
|
| 3061 |
+
},
|
| 3062 |
+
{
|
| 3063 |
+
"epoch": 0.5139870888410698,
|
| 3064 |
+
"grad_norm": 0.3257334260528596,
|
| 3065 |
+
"learning_rate": 1.1264417368505981e-05,
|
| 3066 |
+
"loss": 0.2097,
|
| 3067 |
+
"step": 418
|
| 3068 |
+
},
|
| 3069 |
+
{
|
| 3070 |
+
"epoch": 0.5152167230249001,
|
| 3071 |
+
"grad_norm": 0.37892769382651426,
|
| 3072 |
+
"learning_rate": 1.1221774108241646e-05,
|
| 3073 |
+
"loss": 0.1604,
|
| 3074 |
+
"step": 419
|
| 3075 |
+
},
|
| 3076 |
+
{
|
| 3077 |
+
"epoch": 0.5164463572087304,
|
| 3078 |
+
"grad_norm": 0.424302046031257,
|
| 3079 |
+
"learning_rate": 1.117910828194319e-05,
|
| 3080 |
+
"loss": 0.1472,
|
| 3081 |
+
"step": 420
|
| 3082 |
+
},
|
| 3083 |
+
{
|
| 3084 |
+
"epoch": 0.5176759913925607,
|
| 3085 |
+
"grad_norm": 0.36620159075013065,
|
| 3086 |
+
"learning_rate": 1.1136420677643763e-05,
|
| 3087 |
+
"loss": 0.1656,
|
| 3088 |
+
"step": 421
|
| 3089 |
+
},
|
| 3090 |
+
{
|
| 3091 |
+
"epoch": 0.518905625576391,
|
| 3092 |
+
"grad_norm": 0.3640140474940355,
|
| 3093 |
+
"learning_rate": 1.1093712083778748e-05,
|
| 3094 |
+
"loss": 0.1676,
|
| 3095 |
+
"step": 422
|
| 3096 |
+
},
|
| 3097 |
+
{
|
| 3098 |
+
"epoch": 0.5201352597602213,
|
| 3099 |
+
"grad_norm": 0.5095089795756096,
|
| 3100 |
+
"learning_rate": 1.1050983289171195e-05,
|
| 3101 |
+
"loss": 0.2301,
|
| 3102 |
+
"step": 423
|
| 3103 |
+
},
|
| 3104 |
+
{
|
| 3105 |
+
"epoch": 0.5213648939440516,
|
| 3106 |
+
"grad_norm": 0.3606729676571787,
|
| 3107 |
+
"learning_rate": 1.1008235083017272e-05,
|
| 3108 |
+
"loss": 0.1519,
|
| 3109 |
+
"step": 424
|
| 3110 |
+
},
|
| 3111 |
+
{
|
| 3112 |
+
"epoch": 0.522594528127882,
|
| 3113 |
+
"grad_norm": 0.6045563512444339,
|
| 3114 |
+
"learning_rate": 1.096546825487167e-05,
|
| 3115 |
+
"loss": 0.2366,
|
| 3116 |
+
"step": 425
|
| 3117 |
+
},
|
| 3118 |
+
{
|
| 3119 |
+
"epoch": 0.5238241623117122,
|
| 3120 |
+
"grad_norm": 0.33927673864133545,
|
| 3121 |
+
"learning_rate": 1.092268359463302e-05,
|
| 3122 |
+
"loss": 0.2344,
|
| 3123 |
+
"step": 426
|
| 3124 |
+
},
|
| 3125 |
+
{
|
| 3126 |
+
"epoch": 0.5250537964955426,
|
| 3127 |
+
"grad_norm": 0.42112411326748034,
|
| 3128 |
+
"learning_rate": 1.0879881892529325e-05,
|
| 3129 |
+
"loss": 0.1428,
|
| 3130 |
+
"step": 427
|
| 3131 |
+
},
|
| 3132 |
+
{
|
| 3133 |
+
"epoch": 0.5262834306793729,
|
| 3134 |
+
"grad_norm": 0.5237196748492299,
|
| 3135 |
+
"learning_rate": 1.0837063939103332e-05,
|
| 3136 |
+
"loss": 0.2371,
|
| 3137 |
+
"step": 428
|
| 3138 |
+
},
|
| 3139 |
+
{
|
| 3140 |
+
"epoch": 0.5275130648632032,
|
| 3141 |
+
"grad_norm": 0.45934815233301857,
|
| 3142 |
+
"learning_rate": 1.0794230525197959e-05,
|
| 3143 |
+
"loss": 0.2438,
|
| 3144 |
+
"step": 429
|
| 3145 |
+
},
|
| 3146 |
+
{
|
| 3147 |
+
"epoch": 0.5287426990470335,
|
| 3148 |
+
"grad_norm": 0.6638366111080263,
|
| 3149 |
+
"learning_rate": 1.0751382441941677e-05,
|
| 3150 |
+
"loss": 0.1915,
|
| 3151 |
+
"step": 430
|
| 3152 |
+
},
|
| 3153 |
+
{
|
| 3154 |
+
"epoch": 0.5299723332308638,
|
| 3155 |
+
"grad_norm": 0.4535025325333235,
|
| 3156 |
+
"learning_rate": 1.0708520480733895e-05,
|
| 3157 |
+
"loss": 0.194,
|
| 3158 |
+
"step": 431
|
| 3159 |
+
},
|
| 3160 |
+
{
|
| 3161 |
+
"epoch": 0.5312019674146942,
|
| 3162 |
+
"grad_norm": 0.5223418583566981,
|
| 3163 |
+
"learning_rate": 1.0665645433230345e-05,
|
| 3164 |
+
"loss": 0.2608,
|
| 3165 |
+
"step": 432
|
| 3166 |
+
},
|
| 3167 |
+
{
|
| 3168 |
+
"epoch": 0.5324316015985244,
|
| 3169 |
+
"grad_norm": 0.5251894549889435,
|
| 3170 |
+
"learning_rate": 1.0622758091328469e-05,
|
| 3171 |
+
"loss": 0.1438,
|
| 3172 |
+
"step": 433
|
| 3173 |
+
},
|
| 3174 |
+
{
|
| 3175 |
+
"epoch": 0.5336612357823548,
|
| 3176 |
+
"grad_norm": 0.4790588612821075,
|
| 3177 |
+
"learning_rate": 1.0579859247152774e-05,
|
| 3178 |
+
"loss": 0.1627,
|
| 3179 |
+
"step": 434
|
| 3180 |
+
},
|
| 3181 |
+
{
|
| 3182 |
+
"epoch": 0.5348908699661851,
|
| 3183 |
+
"grad_norm": 0.44616175922026896,
|
| 3184 |
+
"learning_rate": 1.0536949693040224e-05,
|
| 3185 |
+
"loss": 0.2082,
|
| 3186 |
+
"step": 435
|
| 3187 |
+
},
|
| 3188 |
+
{
|
| 3189 |
+
"epoch": 0.5361205041500153,
|
| 3190 |
+
"grad_norm": 0.46980201755709106,
|
| 3191 |
+
"learning_rate": 1.0494030221525582e-05,
|
| 3192 |
+
"loss": 0.2491,
|
| 3193 |
+
"step": 436
|
| 3194 |
+
},
|
| 3195 |
+
{
|
| 3196 |
+
"epoch": 0.5373501383338457,
|
| 3197 |
+
"grad_norm": 0.31327453178418485,
|
| 3198 |
+
"learning_rate": 1.0451101625326798e-05,
|
| 3199 |
+
"loss": 0.1715,
|
| 3200 |
+
"step": 437
|
| 3201 |
+
},
|
| 3202 |
+
{
|
| 3203 |
+
"epoch": 0.538579772517676,
|
| 3204 |
+
"grad_norm": 0.4945036398444153,
|
| 3205 |
+
"learning_rate": 1.0408164697330348e-05,
|
| 3206 |
+
"loss": 0.2311,
|
| 3207 |
+
"step": 438
|
| 3208 |
+
},
|
| 3209 |
+
{
|
| 3210 |
+
"epoch": 0.5398094067015063,
|
| 3211 |
+
"grad_norm": 0.49262428836934985,
|
| 3212 |
+
"learning_rate": 1.0365220230576592e-05,
|
| 3213 |
+
"loss": 0.2812,
|
| 3214 |
+
"step": 439
|
| 3215 |
+
},
|
| 3216 |
+
{
|
| 3217 |
+
"epoch": 0.5410390408853366,
|
| 3218 |
+
"grad_norm": 0.3787467213860296,
|
| 3219 |
+
"learning_rate": 1.0322269018245128e-05,
|
| 3220 |
+
"loss": 0.238,
|
| 3221 |
+
"step": 440
|
| 3222 |
+
},
|
| 3223 |
+
{
|
| 3224 |
+
"epoch": 0.5410390408853366,
|
| 3225 |
+
"eval_accuracy": 0.7914438502673797,
|
| 3226 |
+
"eval_f1": 0.4657534246575342,
|
| 3227 |
+
"eval_loss": 0.4313281178474426,
|
| 3228 |
+
"eval_precision": 0.7391304347826086,
|
| 3229 |
+
"eval_recall": 0.34,
|
| 3230 |
+
"eval_runtime": 22.6212,
|
| 3231 |
+
"eval_samples_per_second": 2.21,
|
| 3232 |
+
"eval_steps_per_second": 0.177,
|
| 3233 |
+
"step": 440
|
| 3234 |
+
},
|
| 3235 |
+
{
|
| 3236 |
+
"epoch": 0.5422686750691669,
|
| 3237 |
+
"grad_norm": 0.37936214504954746,
|
| 3238 |
+
"learning_rate": 1.0279311853640157e-05,
|
| 3239 |
+
"loss": 0.2168,
|
| 3240 |
+
"step": 441
|
| 3241 |
+
},
|
| 3242 |
+
{
|
| 3243 |
+
"epoch": 0.5434983092529972,
|
| 3244 |
+
"grad_norm": 0.3206038847726103,
|
| 3245 |
+
"learning_rate": 1.0236349530175807e-05,
|
| 3246 |
+
"loss": 0.1876,
|
| 3247 |
+
"step": 442
|
| 3248 |
+
},
|
| 3249 |
+
{
|
| 3250 |
+
"epoch": 0.5447279434368275,
|
| 3251 |
+
"grad_norm": 0.36229470962362703,
|
| 3252 |
+
"learning_rate": 1.019338284136149e-05,
|
| 3253 |
+
"loss": 0.1736,
|
| 3254 |
+
"step": 443
|
| 3255 |
+
},
|
| 3256 |
+
{
|
| 3257 |
+
"epoch": 0.5459575776206579,
|
| 3258 |
+
"grad_norm": 0.4075063721479331,
|
| 3259 |
+
"learning_rate": 1.015041258078725e-05,
|
| 3260 |
+
"loss": 0.2283,
|
| 3261 |
+
"step": 444
|
| 3262 |
+
},
|
| 3263 |
+
{
|
| 3264 |
+
"epoch": 0.5471872118044881,
|
| 3265 |
+
"grad_norm": 0.2649302419780058,
|
| 3266 |
+
"learning_rate": 1.0107439542109097e-05,
|
| 3267 |
+
"loss": 0.1342,
|
| 3268 |
+
"step": 445
|
| 3269 |
+
},
|
| 3270 |
+
{
|
| 3271 |
+
"epoch": 0.5484168459883185,
|
| 3272 |
+
"grad_norm": 0.4132655463506092,
|
| 3273 |
+
"learning_rate": 1.0064464519034358e-05,
|
| 3274 |
+
"loss": 0.2227,
|
| 3275 |
+
"step": 446
|
| 3276 |
+
},
|
| 3277 |
+
{
|
| 3278 |
+
"epoch": 0.5496464801721488,
|
| 3279 |
+
"grad_norm": 0.3696111104824441,
|
| 3280 |
+
"learning_rate": 1.0021488305307003e-05,
|
| 3281 |
+
"loss": 0.16,
|
| 3282 |
+
"step": 447
|
| 3283 |
+
},
|
| 3284 |
+
{
|
| 3285 |
+
"epoch": 0.550876114355979,
|
| 3286 |
+
"grad_norm": 0.33812235844778665,
|
| 3287 |
+
"learning_rate": 9.978511694692999e-06,
|
| 3288 |
+
"loss": 0.1787,
|
| 3289 |
+
"step": 448
|
| 3290 |
+
},
|
| 3291 |
+
{
|
| 3292 |
+
"epoch": 0.5521057485398094,
|
| 3293 |
+
"grad_norm": 0.4571938907087761,
|
| 3294 |
+
"learning_rate": 9.935535480965647e-06,
|
| 3295 |
+
"loss": 0.176,
|
| 3296 |
+
"step": 449
|
| 3297 |
+
},
|
| 3298 |
+
{
|
| 3299 |
+
"epoch": 0.5533353827236397,
|
| 3300 |
+
"grad_norm": 0.4297886084823541,
|
| 3301 |
+
"learning_rate": 9.892560457890907e-06,
|
| 3302 |
+
"loss": 0.1822,
|
| 3303 |
+
"step": 450
|
| 3304 |
+
},
|
| 3305 |
+
{
|
| 3306 |
+
"epoch": 0.5545650169074701,
|
| 3307 |
+
"grad_norm": 0.39629426107785415,
|
| 3308 |
+
"learning_rate": 9.849587419212751e-06,
|
| 3309 |
+
"loss": 0.1982,
|
| 3310 |
+
"step": 451
|
| 3311 |
+
},
|
| 3312 |
+
{
|
| 3313 |
+
"epoch": 0.5557946510913003,
|
| 3314 |
+
"grad_norm": 0.4188025429095887,
|
| 3315 |
+
"learning_rate": 9.806617158638515e-06,
|
| 3316 |
+
"loss": 0.263,
|
| 3317 |
+
"step": 452
|
| 3318 |
+
},
|
| 3319 |
+
{
|
| 3320 |
+
"epoch": 0.5570242852751307,
|
| 3321 |
+
"grad_norm": 0.5523505686855087,
|
| 3322 |
+
"learning_rate": 9.763650469824198e-06,
|
| 3323 |
+
"loss": 0.3381,
|
| 3324 |
+
"step": 453
|
| 3325 |
+
},
|
| 3326 |
+
{
|
| 3327 |
+
"epoch": 0.558253919458961,
|
| 3328 |
+
"grad_norm": 0.3507777011017874,
|
| 3329 |
+
"learning_rate": 9.720688146359843e-06,
|
| 3330 |
+
"loss": 0.2102,
|
| 3331 |
+
"step": 454
|
| 3332 |
+
},
|
| 3333 |
+
{
|
| 3334 |
+
"epoch": 0.5594835536427912,
|
| 3335 |
+
"grad_norm": 0.321426765671579,
|
| 3336 |
+
"learning_rate": 9.677730981754875e-06,
|
| 3337 |
+
"loss": 0.1714,
|
| 3338 |
+
"step": 455
|
| 3339 |
+
},
|
| 3340 |
+
{
|
| 3341 |
+
"epoch": 0.5607131878266216,
|
| 3342 |
+
"grad_norm": 0.41323958284891527,
|
| 3343 |
+
"learning_rate": 9.634779769423412e-06,
|
| 3344 |
+
"loss": 0.236,
|
| 3345 |
+
"step": 456
|
| 3346 |
+
},
|
| 3347 |
+
{
|
| 3348 |
+
"epoch": 0.5619428220104519,
|
| 3349 |
+
"grad_norm": 0.49361377740156237,
|
| 3350 |
+
"learning_rate": 9.591835302669657e-06,
|
| 3351 |
+
"loss": 0.1847,
|
| 3352 |
+
"step": 457
|
| 3353 |
+
},
|
| 3354 |
+
{
|
| 3355 |
+
"epoch": 0.5631724561942822,
|
| 3356 |
+
"grad_norm": 0.40610616606906547,
|
| 3357 |
+
"learning_rate": 9.548898374673205e-06,
|
| 3358 |
+
"loss": 0.2124,
|
| 3359 |
+
"step": 458
|
| 3360 |
+
},
|
| 3361 |
+
{
|
| 3362 |
+
"epoch": 0.5644020903781125,
|
| 3363 |
+
"grad_norm": 0.6542907937118401,
|
| 3364 |
+
"learning_rate": 9.505969778474418e-06,
|
| 3365 |
+
"loss": 0.2063,
|
| 3366 |
+
"step": 459
|
| 3367 |
+
},
|
| 3368 |
+
{
|
| 3369 |
+
"epoch": 0.5656317245619428,
|
| 3370 |
+
"grad_norm": 0.34192984136591886,
|
| 3371 |
+
"learning_rate": 9.463050306959782e-06,
|
| 3372 |
+
"loss": 0.1661,
|
| 3373 |
+
"step": 460
|
| 3374 |
+
},
|
| 3375 |
+
{
|
| 3376 |
+
"epoch": 0.5668613587457731,
|
| 3377 |
+
"grad_norm": 0.4336541383620103,
|
| 3378 |
+
"learning_rate": 9.42014075284723e-06,
|
| 3379 |
+
"loss": 0.1666,
|
| 3380 |
+
"step": 461
|
| 3381 |
+
},
|
| 3382 |
+
{
|
| 3383 |
+
"epoch": 0.5680909929296034,
|
| 3384 |
+
"grad_norm": 0.3660193380978002,
|
| 3385 |
+
"learning_rate": 9.377241908671533e-06,
|
| 3386 |
+
"loss": 0.189,
|
| 3387 |
+
"step": 462
|
| 3388 |
+
},
|
| 3389 |
+
{
|
| 3390 |
+
"epoch": 0.5693206271134338,
|
| 3391 |
+
"grad_norm": 0.41494751800096213,
|
| 3392 |
+
"learning_rate": 9.334354566769658e-06,
|
| 3393 |
+
"loss": 0.2119,
|
| 3394 |
+
"step": 463
|
| 3395 |
+
},
|
| 3396 |
+
{
|
| 3397 |
+
"epoch": 0.570550261297264,
|
| 3398 |
+
"grad_norm": 0.5184551716970647,
|
| 3399 |
+
"learning_rate": 9.291479519266108e-06,
|
| 3400 |
+
"loss": 0.2418,
|
| 3401 |
+
"step": 464
|
| 3402 |
+
},
|
| 3403 |
+
{
|
| 3404 |
+
"epoch": 0.5717798954810944,
|
| 3405 |
+
"grad_norm": 0.46046449118174804,
|
| 3406 |
+
"learning_rate": 9.248617558058328e-06,
|
| 3407 |
+
"loss": 0.1998,
|
| 3408 |
+
"step": 465
|
| 3409 |
+
},
|
| 3410 |
+
{
|
| 3411 |
+
"epoch": 0.5730095296649247,
|
| 3412 |
+
"grad_norm": 0.3975308500337465,
|
| 3413 |
+
"learning_rate": 9.205769474802045e-06,
|
| 3414 |
+
"loss": 0.1907,
|
| 3415 |
+
"step": 466
|
| 3416 |
+
},
|
| 3417 |
+
{
|
| 3418 |
+
"epoch": 0.574239163848755,
|
| 3419 |
+
"grad_norm": 0.4237299624678481,
|
| 3420 |
+
"learning_rate": 9.162936060896672e-06,
|
| 3421 |
+
"loss": 0.2087,
|
| 3422 |
+
"step": 467
|
| 3423 |
+
},
|
| 3424 |
+
{
|
| 3425 |
+
"epoch": 0.5754687980325853,
|
| 3426 |
+
"grad_norm": 0.47636428636745337,
|
| 3427 |
+
"learning_rate": 9.12011810747068e-06,
|
| 3428 |
+
"loss": 0.161,
|
| 3429 |
+
"step": 468
|
| 3430 |
+
},
|
| 3431 |
+
{
|
| 3432 |
+
"epoch": 0.5766984322164156,
|
| 3433 |
+
"grad_norm": 0.5257948156557456,
|
| 3434 |
+
"learning_rate": 9.07731640536698e-06,
|
| 3435 |
+
"loss": 0.3091,
|
| 3436 |
+
"step": 469
|
| 3437 |
+
},
|
| 3438 |
+
{
|
| 3439 |
+
"epoch": 0.577928066400246,
|
| 3440 |
+
"grad_norm": 0.4479489865327413,
|
| 3441 |
+
"learning_rate": 9.034531745128334e-06,
|
| 3442 |
+
"loss": 0.2383,
|
| 3443 |
+
"step": 470
|
| 3444 |
+
},
|
| 3445 |
+
{
|
| 3446 |
+
"epoch": 0.5791577005840762,
|
| 3447 |
+
"grad_norm": 0.40206612139181935,
|
| 3448 |
+
"learning_rate": 8.991764916982731e-06,
|
| 3449 |
+
"loss": 0.2534,
|
| 3450 |
+
"step": 471
|
| 3451 |
+
},
|
| 3452 |
+
{
|
| 3453 |
+
"epoch": 0.5803873347679065,
|
| 3454 |
+
"grad_norm": 0.3870155682325831,
|
| 3455 |
+
"learning_rate": 8.949016710828808e-06,
|
| 3456 |
+
"loss": 0.1743,
|
| 3457 |
+
"step": 472
|
| 3458 |
+
},
|
| 3459 |
+
{
|
| 3460 |
+
"epoch": 0.5816169689517369,
|
| 3461 |
+
"grad_norm": 0.35857102895539483,
|
| 3462 |
+
"learning_rate": 8.906287916221259e-06,
|
| 3463 |
+
"loss": 0.1541,
|
| 3464 |
+
"step": 473
|
| 3465 |
+
},
|
| 3466 |
+
{
|
| 3467 |
+
"epoch": 0.5828466031355671,
|
| 3468 |
+
"grad_norm": 0.3297878518016191,
|
| 3469 |
+
"learning_rate": 8.863579322356242e-06,
|
| 3470 |
+
"loss": 0.184,
|
| 3471 |
+
"step": 474
|
| 3472 |
+
},
|
| 3473 |
+
{
|
| 3474 |
+
"epoch": 0.5840762373193975,
|
| 3475 |
+
"grad_norm": 0.2983994696232135,
|
| 3476 |
+
"learning_rate": 8.820891718056815e-06,
|
| 3477 |
+
"loss": 0.1836,
|
| 3478 |
+
"step": 475
|
| 3479 |
+
},
|
| 3480 |
+
{
|
| 3481 |
+
"epoch": 0.5853058715032278,
|
| 3482 |
+
"grad_norm": 0.40664636398131304,
|
| 3483 |
+
"learning_rate": 8.77822589175836e-06,
|
| 3484 |
+
"loss": 0.1594,
|
| 3485 |
+
"step": 476
|
| 3486 |
+
},
|
| 3487 |
+
{
|
| 3488 |
+
"epoch": 0.5865355056870581,
|
| 3489 |
+
"grad_norm": 0.37741982285626524,
|
| 3490 |
+
"learning_rate": 8.73558263149402e-06,
|
| 3491 |
+
"loss": 0.236,
|
| 3492 |
+
"step": 477
|
| 3493 |
+
},
|
| 3494 |
+
{
|
| 3495 |
+
"epoch": 0.5877651398708884,
|
| 3496 |
+
"grad_norm": 0.4536143968466746,
|
| 3497 |
+
"learning_rate": 8.692962724880148e-06,
|
| 3498 |
+
"loss": 0.1882,
|
| 3499 |
+
"step": 478
|
| 3500 |
+
},
|
| 3501 |
+
{
|
| 3502 |
+
"epoch": 0.5889947740547187,
|
| 3503 |
+
"grad_norm": 0.3713741062634612,
|
| 3504 |
+
"learning_rate": 8.650366959101757e-06,
|
| 3505 |
+
"loss": 0.2143,
|
| 3506 |
+
"step": 479
|
| 3507 |
+
},
|
| 3508 |
+
{
|
| 3509 |
+
"epoch": 0.590224408238549,
|
| 3510 |
+
"grad_norm": 0.27112035186414296,
|
| 3511 |
+
"learning_rate": 8.607796120897978e-06,
|
| 3512 |
+
"loss": 0.1448,
|
| 3513 |
+
"step": 480
|
| 3514 |
+
},
|
| 3515 |
+
{
|
| 3516 |
+
"epoch": 0.590224408238549,
|
| 3517 |
+
"eval_accuracy": 0.8128342245989305,
|
| 3518 |
+
"eval_f1": 0.5454545454545454,
|
| 3519 |
+
"eval_loss": 0.41609373688697815,
|
| 3520 |
+
"eval_precision": 0.7777777777777778,
|
| 3521 |
+
"eval_recall": 0.42,
|
| 3522 |
+
"eval_runtime": 23.0163,
|
| 3523 |
+
"eval_samples_per_second": 2.172,
|
| 3524 |
+
"eval_steps_per_second": 0.174,
|
| 3525 |
+
"step": 480
|
| 3526 |
+
},
|
| 3527 |
+
{
|
| 3528 |
+
"epoch": 0.5914540424223793,
|
| 3529 |
+
"grad_norm": 0.3731696583376688,
|
| 3530 |
+
"learning_rate": 8.565250996547538e-06,
|
| 3531 |
+
"loss": 0.1521,
|
| 3532 |
+
"step": 481
|
| 3533 |
+
},
|
| 3534 |
+
{
|
| 3535 |
+
"epoch": 0.5926836766062097,
|
| 3536 |
+
"grad_norm": 0.3715423359630779,
|
| 3537 |
+
"learning_rate": 8.522732371854228e-06,
|
| 3538 |
+
"loss": 0.1561,
|
| 3539 |
+
"step": 482
|
| 3540 |
+
},
|
| 3541 |
+
{
|
| 3542 |
+
"epoch": 0.59391331079004,
|
| 3543 |
+
"grad_norm": 0.4445264123453298,
|
| 3544 |
+
"learning_rate": 8.480241032132394e-06,
|
| 3545 |
+
"loss": 0.1772,
|
| 3546 |
+
"step": 483
|
| 3547 |
+
},
|
| 3548 |
+
{
|
| 3549 |
+
"epoch": 0.5951429449738703,
|
| 3550 |
+
"grad_norm": 0.417726015870805,
|
| 3551 |
+
"learning_rate": 8.437777762192434e-06,
|
| 3552 |
+
"loss": 0.2099,
|
| 3553 |
+
"step": 484
|
| 3554 |
+
},
|
| 3555 |
+
{
|
| 3556 |
+
"epoch": 0.5963725791577006,
|
| 3557 |
+
"grad_norm": 0.5443603401932148,
|
| 3558 |
+
"learning_rate": 8.395343346326295e-06,
|
| 3559 |
+
"loss": 0.2194,
|
| 3560 |
+
"step": 485
|
| 3561 |
+
},
|
| 3562 |
+
{
|
| 3563 |
+
"epoch": 0.5976022133415309,
|
| 3564 |
+
"grad_norm": 0.357174859123756,
|
| 3565 |
+
"learning_rate": 8.352938568293e-06,
|
| 3566 |
+
"loss": 0.1994,
|
| 3567 |
+
"step": 486
|
| 3568 |
+
},
|
| 3569 |
+
{
|
| 3570 |
+
"epoch": 0.5988318475253612,
|
| 3571 |
+
"grad_norm": 0.4122807392874181,
|
| 3572 |
+
"learning_rate": 8.310564211304159e-06,
|
| 3573 |
+
"loss": 0.2326,
|
| 3574 |
+
"step": 487
|
| 3575 |
+
},
|
| 3576 |
+
{
|
| 3577 |
+
"epoch": 0.6000614817091915,
|
| 3578 |
+
"grad_norm": 0.3030477984709631,
|
| 3579 |
+
"learning_rate": 8.268221058009506e-06,
|
| 3580 |
+
"loss": 0.1678,
|
| 3581 |
+
"step": 488
|
| 3582 |
+
},
|
| 3583 |
+
{
|
| 3584 |
+
"epoch": 0.6012911158930219,
|
| 3585 |
+
"grad_norm": 0.32220394621350373,
|
| 3586 |
+
"learning_rate": 8.225909890482456e-06,
|
| 3587 |
+
"loss": 0.1442,
|
| 3588 |
+
"step": 489
|
| 3589 |
+
},
|
| 3590 |
+
{
|
| 3591 |
+
"epoch": 0.6025207500768521,
|
| 3592 |
+
"grad_norm": 0.4724721543903087,
|
| 3593 |
+
"learning_rate": 8.183631490205636e-06,
|
| 3594 |
+
"loss": 0.1962,
|
| 3595 |
+
"step": 490
|
| 3596 |
+
},
|
| 3597 |
+
{
|
| 3598 |
+
"epoch": 0.6037503842606824,
|
| 3599 |
+
"grad_norm": 0.3600147393401425,
|
| 3600 |
+
"learning_rate": 8.141386638056482e-06,
|
| 3601 |
+
"loss": 0.1876,
|
| 3602 |
+
"step": 491
|
| 3603 |
+
},
|
| 3604 |
+
{
|
| 3605 |
+
"epoch": 0.6049800184445128,
|
| 3606 |
+
"grad_norm": 0.4000861620375486,
|
| 3607 |
+
"learning_rate": 8.09917611429279e-06,
|
| 3608 |
+
"loss": 0.2202,
|
| 3609 |
+
"step": 492
|
| 3610 |
+
},
|
| 3611 |
+
{
|
| 3612 |
+
"epoch": 0.606209652628343,
|
| 3613 |
+
"grad_norm": 0.3772045360701519,
|
| 3614 |
+
"learning_rate": 8.057000698538311e-06,
|
| 3615 |
+
"loss": 0.1606,
|
| 3616 |
+
"step": 493
|
| 3617 |
+
},
|
| 3618 |
+
{
|
| 3619 |
+
"epoch": 0.6074392868121734,
|
| 3620 |
+
"grad_norm": 0.37207256879831996,
|
| 3621 |
+
"learning_rate": 8.014861169768362e-06,
|
| 3622 |
+
"loss": 0.2028,
|
| 3623 |
+
"step": 494
|
| 3624 |
+
},
|
| 3625 |
+
{
|
| 3626 |
+
"epoch": 0.6086689209960037,
|
| 3627 |
+
"grad_norm": 0.3479563238322552,
|
| 3628 |
+
"learning_rate": 7.972758306295436e-06,
|
| 3629 |
+
"loss": 0.1574,
|
| 3630 |
+
"step": 495
|
| 3631 |
+
},
|
| 3632 |
+
{
|
| 3633 |
+
"epoch": 0.609898555179834,
|
| 3634 |
+
"grad_norm": 0.4796467447401053,
|
| 3635 |
+
"learning_rate": 7.930692885754806e-06,
|
| 3636 |
+
"loss": 0.224,
|
| 3637 |
+
"step": 496
|
| 3638 |
+
},
|
| 3639 |
+
{
|
| 3640 |
+
"epoch": 0.6111281893636643,
|
| 3641 |
+
"grad_norm": 0.3477407213433725,
|
| 3642 |
+
"learning_rate": 7.888665685090194e-06,
|
| 3643 |
+
"loss": 0.2333,
|
| 3644 |
+
"step": 497
|
| 3645 |
+
},
|
| 3646 |
+
{
|
| 3647 |
+
"epoch": 0.6123578235474946,
|
| 3648 |
+
"grad_norm": 0.3793733407594908,
|
| 3649 |
+
"learning_rate": 7.846677480539392e-06,
|
| 3650 |
+
"loss": 0.2162,
|
| 3651 |
+
"step": 498
|
| 3652 |
+
},
|
| 3653 |
+
{
|
| 3654 |
+
"epoch": 0.613587457731325,
|
| 3655 |
+
"grad_norm": 0.3805341456576239,
|
| 3656 |
+
"learning_rate": 7.80472904761995e-06,
|
| 3657 |
+
"loss": 0.141,
|
| 3658 |
+
"step": 499
|
| 3659 |
+
},
|
| 3660 |
+
{
|
| 3661 |
+
"epoch": 0.6148170919151552,
|
| 3662 |
+
"grad_norm": 0.29991453914514893,
|
| 3663 |
+
"learning_rate": 7.762821161114834e-06,
|
| 3664 |
+
"loss": 0.1767,
|
| 3665 |
+
"step": 500
|
| 3666 |
}
|
| 3667 |
],
|
| 3668 |
"logging_steps": 1,
|
|
|
|
| 3682 |
"attributes": {}
|
| 3683 |
}
|
| 3684 |
},
|
| 3685 |
+
"total_flos": 491155566985216.0,
|
| 3686 |
"train_batch_size": 4,
|
| 3687 |
"trial_name": null,
|
| 3688 |
"trial_params": null
|