Upload folder using huggingface_hub
Browse files- checkpoints-v4.3+/checkpoint-14336/ema.safetensors +3 -0
- checkpoints-v4.3+/checkpoint-14336/model.safetensors +3 -0
- checkpoints-v4.3+/checkpoint-14336/optimizer.pt +3 -0
- checkpoints-v4.3+/checkpoint-14336/rng_state.pth +3 -0
- checkpoints-v4.3+/checkpoint-14336/scaler.pt +3 -0
- checkpoints-v4.3+/checkpoint-14336/scheduler.pt +3 -0
- checkpoints-v4.3+/checkpoint-14336/trainer_state.json +804 -0
- checkpoints-v4.3+/checkpoint-14336/training_args.bin +3 -0
checkpoints-v4.3+/checkpoint-14336/ema.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:35a46f33817451f78451d0ed373fc5d437a4ae6b6761dd1284353c5d42ceda68
|
| 3 |
+
size 4843064
|
checkpoints-v4.3+/checkpoint-14336/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e6452f5655f35ea3aa8bc8c01221cb06a14c68878dbfbf1f38bb5fbaaa72f4e2
|
| 3 |
+
size 4843096
|
checkpoints-v4.3+/checkpoint-14336/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2a7e8d9bdf812acd7cb65d919bf38450210a28d8802137dc40e919a0c5d2a028
|
| 3 |
+
size 9762827
|
checkpoints-v4.3+/checkpoint-14336/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f39c33e67e77d34ac55da0f13f72e7b8df48c27a34ba96f5523163c6029c0cef
|
| 3 |
+
size 14645
|
checkpoints-v4.3+/checkpoint-14336/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e8dee4011adae7ae78189be2be21a9f25228e843770b17776feb738b8e6025d5
|
| 3 |
+
size 1383
|
checkpoints-v4.3+/checkpoint-14336/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:43b80f7b54ea40edb7a4cdcd3092f35ad902ad030ad30567ab9fdfb8020c9dc4
|
| 3 |
+
size 1465
|
checkpoints-v4.3+/checkpoint-14336/trainer_state.json
ADDED
|
@@ -0,0 +1,804 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.6621403168444876,
|
| 6 |
+
"eval_steps": 1024,
|
| 7 |
+
"global_step": 14336,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.011823934229365849,
|
| 14 |
+
"grad_norm": 5.310993194580078,
|
| 15 |
+
"learning_rate": 0.000498046875,
|
| 16 |
+
"loss": 22.2890625,
|
| 17 |
+
"step": 256
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.023647868458731697,
|
| 21 |
+
"grad_norm": 7.661691188812256,
|
| 22 |
+
"learning_rate": 0.000998046875,
|
| 23 |
+
"loss": 12.93233871459961,
|
| 24 |
+
"step": 512
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.03547180268809755,
|
| 28 |
+
"grad_norm": 2.532165050506592,
|
| 29 |
+
"learning_rate": 0.000999640996023194,
|
| 30 |
+
"loss": 9.235491752624512,
|
| 31 |
+
"step": 768
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.047295736917463395,
|
| 35 |
+
"grad_norm": 3.695277214050293,
|
| 36 |
+
"learning_rate": 0.0009985588674043958,
|
| 37 |
+
"loss": 8.26991081237793,
|
| 38 |
+
"step": 1024
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.047295736917463395,
|
| 42 |
+
"eval_ce_loss": 6.770938226621445,
|
| 43 |
+
"eval_cos_loss": 0.5271919946011887,
|
| 44 |
+
"eval_loss": 8.123252781558799,
|
| 45 |
+
"eval_mse_loss": 1.3523145494940074,
|
| 46 |
+
"flow/cos_sim": 0.47280800812048457,
|
| 47 |
+
"flow/improvement_ratio": 0.9578989227068478,
|
| 48 |
+
"flow/mag_ratio_mean": 0.514818922025428,
|
| 49 |
+
"flow/mag_ratio_std": 0.19883525225261575,
|
| 50 |
+
"step": 1024
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.047295736917463395,
|
| 54 |
+
"eval_ce_loss": 6.770938226621445,
|
| 55 |
+
"eval_cos_loss": 0.5271919946011887,
|
| 56 |
+
"eval_loss": 8.123252781558799,
|
| 57 |
+
"eval_mse_loss": 1.3523145494940074,
|
| 58 |
+
"eval_runtime": 64.4558,
|
| 59 |
+
"eval_samples_per_second": 434.297,
|
| 60 |
+
"eval_steps_per_second": 6.795,
|
| 61 |
+
"flow/cos_sim": 0.47280800812048457,
|
| 62 |
+
"flow/improvement_ratio": 0.9578989227068478,
|
| 63 |
+
"flow/mag_ratio_mean": 0.514818922025428,
|
| 64 |
+
"flow/mag_ratio_std": 0.19883525225261575,
|
| 65 |
+
"step": 1024
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.05911967114682925,
|
| 69 |
+
"grad_norm": 2.314648389816284,
|
| 70 |
+
"learning_rate": 0.0009967551747861387,
|
| 71 |
+
"loss": 8.022788047790527,
|
| 72 |
+
"step": 1280
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.0709436053761951,
|
| 76 |
+
"grad_norm": 1.662047266960144,
|
| 77 |
+
"learning_rate": 0.000994232528651847,
|
| 78 |
+
"loss": 7.858725070953369,
|
| 79 |
+
"step": 1536
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.08276753960556095,
|
| 83 |
+
"grad_norm": 1.137885570526123,
|
| 84 |
+
"learning_rate": 0.0009909945800260092,
|
| 85 |
+
"loss": 7.710772514343262,
|
| 86 |
+
"step": 1792
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.09459147383492679,
|
| 90 |
+
"grad_norm": 1.100211262702942,
|
| 91 |
+
"learning_rate": 0.0009870460151900522,
|
| 92 |
+
"loss": 7.598612308502197,
|
| 93 |
+
"step": 2048
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.09459147383492679,
|
| 97 |
+
"eval_ce_loss": 6.6824955722512716,
|
| 98 |
+
"eval_cos_loss": 0.39051488293632525,
|
| 99 |
+
"eval_loss": 7.547230463594063,
|
| 100 |
+
"eval_mse_loss": 0.864734881272599,
|
| 101 |
+
"flow/cos_sim": 0.6094851142739597,
|
| 102 |
+
"flow/improvement_ratio": 0.9694367877152412,
|
| 103 |
+
"flow/mag_ratio_mean": 0.6239872499960198,
|
| 104 |
+
"flow/mag_ratio_std": 0.20368898976339053,
|
| 105 |
+
"step": 2048
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.09459147383492679,
|
| 109 |
+
"eval_ce_loss": 6.6824955722512716,
|
| 110 |
+
"eval_cos_loss": 0.39051488293632525,
|
| 111 |
+
"eval_loss": 7.547230463594063,
|
| 112 |
+
"eval_mse_loss": 0.864734881272599,
|
| 113 |
+
"eval_runtime": 64.6729,
|
| 114 |
+
"eval_samples_per_second": 432.84,
|
| 115 |
+
"eval_steps_per_second": 6.773,
|
| 116 |
+
"flow/cos_sim": 0.6094851142739597,
|
| 117 |
+
"flow/improvement_ratio": 0.9694367877152412,
|
| 118 |
+
"flow/mag_ratio_mean": 0.6239872499960198,
|
| 119 |
+
"flow/mag_ratio_std": 0.20368898976339053,
|
| 120 |
+
"step": 2048
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 0.10641540806429264,
|
| 124 |
+
"grad_norm": 1.3531696796417236,
|
| 125 |
+
"learning_rate": 0.0009823925488998885,
|
| 126 |
+
"loss": 7.5045647621154785,
|
| 127 |
+
"step": 2304
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 0.1182393422936585,
|
| 131 |
+
"grad_norm": 0.7993828654289246,
|
| 132 |
+
"learning_rate": 0.0009770409161149525,
|
| 133 |
+
"loss": 7.426115989685059,
|
| 134 |
+
"step": 2560
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.13006327652302435,
|
| 138 |
+
"grad_norm": 0.678104043006897,
|
| 139 |
+
"learning_rate": 0.0009709988622506973,
|
| 140 |
+
"loss": 7.355522632598877,
|
| 141 |
+
"step": 2816
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.1418872107523902,
|
| 145 |
+
"grad_norm": 0.5943172574043274,
|
| 146 |
+
"learning_rate": 0.000964275131968659,
|
| 147 |
+
"loss": 7.290462493896484,
|
| 148 |
+
"step": 3072
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 0.1418872107523902,
|
| 152 |
+
"eval_ce_loss": 6.662393192178038,
|
| 153 |
+
"eval_cos_loss": 0.2890751386001774,
|
| 154 |
+
"eval_loss": 7.268133973422116,
|
| 155 |
+
"eval_mse_loss": 0.6057407744398945,
|
| 156 |
+
"flow/cos_sim": 0.710924859188463,
|
| 157 |
+
"flow/improvement_ratio": 0.9773364528401257,
|
| 158 |
+
"flow/mag_ratio_mean": 0.724724873288037,
|
| 159 |
+
"flow/mag_ratio_std": 0.22310954716788034,
|
| 160 |
+
"step": 3072
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 0.1418872107523902,
|
| 164 |
+
"eval_ce_loss": 6.662393192178038,
|
| 165 |
+
"eval_cos_loss": 0.2890751386001774,
|
| 166 |
+
"eval_loss": 7.268133973422116,
|
| 167 |
+
"eval_mse_loss": 0.6057407744398945,
|
| 168 |
+
"eval_runtime": 64.5435,
|
| 169 |
+
"eval_samples_per_second": 433.708,
|
| 170 |
+
"eval_steps_per_second": 6.786,
|
| 171 |
+
"flow/cos_sim": 0.710924859188463,
|
| 172 |
+
"flow/improvement_ratio": 0.9773364528401257,
|
| 173 |
+
"flow/mag_ratio_mean": 0.724724873288037,
|
| 174 |
+
"flow/mag_ratio_std": 0.22310954716788034,
|
| 175 |
+
"step": 3072
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 0.15371114498175603,
|
| 179 |
+
"grad_norm": 1.2431687116622925,
|
| 180 |
+
"learning_rate": 0.0009568794565203123,
|
| 181 |
+
"loss": 7.288367748260498,
|
| 182 |
+
"step": 3328
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.1655350792111219,
|
| 186 |
+
"grad_norm": 0.9812341928482056,
|
| 187 |
+
"learning_rate": 0.0009488225396630347,
|
| 188 |
+
"loss": 7.193031311035156,
|
| 189 |
+
"step": 3584
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.17735901344048774,
|
| 193 |
+
"grad_norm": 0.46874845027923584,
|
| 194 |
+
"learning_rate": 0.0009401160421685646,
|
| 195 |
+
"loss": 7.153883457183838,
|
| 196 |
+
"step": 3840
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 0.18918294766985358,
|
| 200 |
+
"grad_norm": 0.47678303718566895,
|
| 201 |
+
"learning_rate": 0.0009307725649463714,
|
| 202 |
+
"loss": 7.12277889251709,
|
| 203 |
+
"step": 4096
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 0.18918294766985358,
|
| 207 |
+
"eval_ce_loss": 6.650159499416612,
|
| 208 |
+
"eval_cos_loss": 0.22240412983720162,
|
| 209 |
+
"eval_loss": 7.105518678551939,
|
| 210 |
+
"eval_mse_loss": 0.4553591773662393,
|
| 211 |
+
"flow/cos_sim": 0.7775958679854598,
|
| 212 |
+
"flow/improvement_ratio": 0.9806619151814343,
|
| 213 |
+
"flow/mag_ratio_mean": 0.7837155504858113,
|
| 214 |
+
"flow/mag_ratio_std": 0.21672996865015595,
|
| 215 |
+
"step": 4096
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 0.18918294766985358,
|
| 219 |
+
"eval_ce_loss": 6.650159499416612,
|
| 220 |
+
"eval_cos_loss": 0.22240412983720162,
|
| 221 |
+
"eval_loss": 7.105518678551939,
|
| 222 |
+
"eval_mse_loss": 0.4553591773662393,
|
| 223 |
+
"eval_runtime": 63.9268,
|
| 224 |
+
"eval_samples_per_second": 437.891,
|
| 225 |
+
"eval_steps_per_second": 6.852,
|
| 226 |
+
"flow/cos_sim": 0.7775958679854598,
|
| 227 |
+
"flow/improvement_ratio": 0.9806619151814343,
|
| 228 |
+
"flow/mag_ratio_mean": 0.7837155504858113,
|
| 229 |
+
"flow/mag_ratio_std": 0.21672996865015595,
|
| 230 |
+
"step": 4096
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.20100688189921945,
|
| 234 |
+
"grad_norm": 0.6425520181655884,
|
| 235 |
+
"learning_rate": 0.0009208056308063659,
|
| 236 |
+
"loss": 7.0932536125183105,
|
| 237 |
+
"step": 4352
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.2128308161285853,
|
| 241 |
+
"grad_norm": 1.3422183990478516,
|
| 242 |
+
"learning_rate": 0.0009102296648873445,
|
| 243 |
+
"loss": 7.0613112449646,
|
| 244 |
+
"step": 4608
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 0.22465475035795113,
|
| 248 |
+
"grad_norm": 0.5829691290855408,
|
| 249 |
+
"learning_rate": 0.0008990599737794927,
|
| 250 |
+
"loss": 7.035637855529785,
|
| 251 |
+
"step": 4864
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 0.236478684587317,
|
| 255 |
+
"grad_norm": 0.5854812860488892,
|
| 256 |
+
"learning_rate": 0.0008873127233711644,
|
| 257 |
+
"loss": 7.0129170417785645,
|
| 258 |
+
"step": 5120
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 0.236478684587317,
|
| 262 |
+
"eval_ce_loss": 6.647287761784035,
|
| 263 |
+
"eval_cos_loss": 0.1774802339267513,
|
| 264 |
+
"eval_loss": 7.006625726342746,
|
| 265 |
+
"eval_mse_loss": 0.3593379625174553,
|
| 266 |
+
"flow/cos_sim": 0.8225197658691232,
|
| 267 |
+
"flow/improvement_ratio": 0.9842402499560352,
|
| 268 |
+
"flow/mag_ratio_mean": 0.8389692129609791,
|
| 269 |
+
"flow/mag_ratio_std": 0.2145905472576346,
|
| 270 |
+
"step": 5120
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 0.236478684587317,
|
| 274 |
+
"eval_ce_loss": 6.647287761784035,
|
| 275 |
+
"eval_cos_loss": 0.1774802339267513,
|
| 276 |
+
"eval_loss": 7.006625726342746,
|
| 277 |
+
"eval_mse_loss": 0.3593379625174553,
|
| 278 |
+
"eval_runtime": 65.6903,
|
| 279 |
+
"eval_samples_per_second": 426.136,
|
| 280 |
+
"eval_steps_per_second": 6.668,
|
| 281 |
+
"flow/cos_sim": 0.8225197658691232,
|
| 282 |
+
"flow/improvement_ratio": 0.9842402499560352,
|
| 283 |
+
"flow/mag_ratio_mean": 0.8389692129609791,
|
| 284 |
+
"flow/mag_ratio_std": 0.2145905472576346,
|
| 285 |
+
"step": 5120
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.24830261881668284,
|
| 289 |
+
"grad_norm": 0.7321183085441589,
|
| 290 |
+
"learning_rate": 0.0008750049154520011,
|
| 291 |
+
"loss": 6.990157604217529,
|
| 292 |
+
"step": 5376
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 0.2601265530460487,
|
| 296 |
+
"grad_norm": 0.5175312757492065,
|
| 297 |
+
"learning_rate": 0.0008621543631062487,
|
| 298 |
+
"loss": 6.97371244430542,
|
| 299 |
+
"step": 5632
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 0.27195048727541454,
|
| 303 |
+
"grad_norm": 0.7752848863601685,
|
| 304 |
+
"learning_rate": 0.0008487796649318904,
|
| 305 |
+
"loss": 6.963197708129883,
|
| 306 |
+
"step": 5888
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 0.2837744215047804,
|
| 310 |
+
"grad_norm": 0.5458380579948425,
|
| 311 |
+
"learning_rate": 0.0008349001781229053,
|
| 312 |
+
"loss": 6.956610679626465,
|
| 313 |
+
"step": 6144
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 0.2837744215047804,
|
| 317 |
+
"eval_ce_loss": 6.64574583806948,
|
| 318 |
+
"eval_cos_loss": 0.14618989437410276,
|
| 319 |
+
"eval_loss": 6.940428107840829,
|
| 320 |
+
"eval_mse_loss": 0.2946822723569391,
|
| 321 |
+
"flow/cos_sim": 0.8538101044691861,
|
| 322 |
+
"flow/improvement_ratio": 0.9873136728321581,
|
| 323 |
+
"flow/mag_ratio_mean": 0.8605127470678391,
|
| 324 |
+
"flow/mag_ratio_std": 0.1925806637223997,
|
| 325 |
+
"step": 6144
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.2837744215047804,
|
| 329 |
+
"eval_ce_loss": 6.64574583806948,
|
| 330 |
+
"eval_cos_loss": 0.14618989437410276,
|
| 331 |
+
"eval_loss": 6.940428107840829,
|
| 332 |
+
"eval_mse_loss": 0.2946822723569391,
|
| 333 |
+
"eval_runtime": 65.5471,
|
| 334 |
+
"eval_samples_per_second": 427.067,
|
| 335 |
+
"eval_steps_per_second": 6.682,
|
| 336 |
+
"flow/cos_sim": 0.8538101044691861,
|
| 337 |
+
"flow/improvement_ratio": 0.9873136728321581,
|
| 338 |
+
"flow/mag_ratio_mean": 0.8605127470678391,
|
| 339 |
+
"flow/mag_ratio_std": 0.1925806637223997,
|
| 340 |
+
"step": 6144
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 0.2955983557341462,
|
| 344 |
+
"grad_norm": 0.437857985496521,
|
| 345 |
+
"learning_rate": 0.0008205359904536107,
|
| 346 |
+
"loss": 6.932642936706543,
|
| 347 |
+
"step": 6400
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 0.30742228996351206,
|
| 351 |
+
"grad_norm": 0.6009621024131775,
|
| 352 |
+
"learning_rate": 0.0008057078912056363,
|
| 353 |
+
"loss": 6.932676315307617,
|
| 354 |
+
"step": 6656
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 0.3192462241928779,
|
| 358 |
+
"grad_norm": 0.4097861051559448,
|
| 359 |
+
"learning_rate": 0.0007904373410796086,
|
| 360 |
+
"loss": 6.922003746032715,
|
| 361 |
+
"step": 6912
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 0.3310701584222438,
|
| 365 |
+
"grad_norm": 0.54083251953125,
|
| 366 |
+
"learning_rate": 0.0007747464411350876,
|
| 367 |
+
"loss": 6.903475284576416,
|
| 368 |
+
"step": 7168
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.3310701584222438,
|
| 372 |
+
"eval_ce_loss": 6.644999326636258,
|
| 373 |
+
"eval_cos_loss": 0.12948155590252244,
|
| 374 |
+
"eval_loss": 6.905599483071941,
|
| 375 |
+
"eval_mse_loss": 0.26060015364596834,
|
| 376 |
+
"flow/cos_sim": 0.87051844270262,
|
| 377 |
+
"flow/improvement_ratio": 0.9893111555271497,
|
| 378 |
+
"flow/mag_ratio_mean": 0.8712390643548747,
|
| 379 |
+
"flow/mag_ratio_std": 0.19501862627322272,
|
| 380 |
+
"step": 7168
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.3310701584222438,
|
| 384 |
+
"eval_ce_loss": 6.644999326636258,
|
| 385 |
+
"eval_cos_loss": 0.12948155590252244,
|
| 386 |
+
"eval_loss": 6.905599483071941,
|
| 387 |
+
"eval_mse_loss": 0.26060015364596834,
|
| 388 |
+
"eval_runtime": 64.508,
|
| 389 |
+
"eval_samples_per_second": 433.946,
|
| 390 |
+
"eval_steps_per_second": 6.79,
|
| 391 |
+
"flow/cos_sim": 0.87051844270262,
|
| 392 |
+
"flow/improvement_ratio": 0.9893111555271497,
|
| 393 |
+
"flow/mag_ratio_mean": 0.8712390643548747,
|
| 394 |
+
"flow/mag_ratio_std": 0.19501862627322272,
|
| 395 |
+
"step": 7168
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.34289409265160964,
|
| 399 |
+
"grad_norm": 0.5095380544662476,
|
| 400 |
+
"learning_rate": 0.000758657900803716,
|
| 401 |
+
"loss": 6.9005961418151855,
|
| 402 |
+
"step": 7424
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 0.3547180268809755,
|
| 406 |
+
"grad_norm": 0.341415673494339,
|
| 407 |
+
"learning_rate": 0.000742195005021869,
|
| 408 |
+
"loss": 6.8952412605285645,
|
| 409 |
+
"step": 7680
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 0.3665419611103413,
|
| 413 |
+
"grad_norm": 0.2902733385562897,
|
| 414 |
+
"learning_rate": 0.0007253815805303786,
|
| 415 |
+
"loss": 6.8849945068359375,
|
| 416 |
+
"step": 7936
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.37836589533970716,
|
| 420 |
+
"grad_norm": 0.5425611138343811,
|
| 421 |
+
"learning_rate": 0.0007082419613901028,
|
| 422 |
+
"loss": 6.8753275871276855,
|
| 423 |
+
"step": 8192
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.37836589533970716,
|
| 427 |
+
"eval_ce_loss": 6.643324745300154,
|
| 428 |
+
"eval_cos_loss": 0.11695763337761844,
|
| 429 |
+
"eval_loss": 6.877702439756698,
|
| 430 |
+
"eval_mse_loss": 0.23437770422054752,
|
| 431 |
+
"flow/cos_sim": 0.8830423678981659,
|
| 432 |
+
"flow/improvement_ratio": 0.9898773733883688,
|
| 433 |
+
"flow/mag_ratio_mean": 0.8962584248144333,
|
| 434 |
+
"flow/mag_ratio_std": 0.18067260168186605,
|
| 435 |
+
"step": 8192
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.37836589533970716,
|
| 439 |
+
"eval_ce_loss": 6.643324745300154,
|
| 440 |
+
"eval_cos_loss": 0.11695763337761844,
|
| 441 |
+
"eval_loss": 6.877702439756698,
|
| 442 |
+
"eval_mse_loss": 0.23437770422054752,
|
| 443 |
+
"eval_runtime": 64.3325,
|
| 444 |
+
"eval_samples_per_second": 435.13,
|
| 445 |
+
"eval_steps_per_second": 6.808,
|
| 446 |
+
"flow/cos_sim": 0.8830423678981659,
|
| 447 |
+
"flow/improvement_ratio": 0.9898773733883688,
|
| 448 |
+
"flow/mag_ratio_mean": 0.8962584248144333,
|
| 449 |
+
"flow/mag_ratio_std": 0.18067260168186605,
|
| 450 |
+
"step": 8192
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.390189829569073,
|
| 454 |
+
"grad_norm": 0.7520615458488464,
|
| 455 |
+
"learning_rate": 0.0006908009537632514,
|
| 456 |
+
"loss": 6.879398345947266,
|
| 457 |
+
"step": 8448
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.4020137637984389,
|
| 461 |
+
"grad_norm": 0.5208486914634705,
|
| 462 |
+
"learning_rate": 0.0006730838000114403,
|
| 463 |
+
"loss": 6.873098373413086,
|
| 464 |
+
"step": 8704
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.41383769802780473,
|
| 468 |
+
"grad_norm": 0.43452462553977966,
|
| 469 |
+
"learning_rate": 0.0006551161421624341,
|
| 470 |
+
"loss": 6.862982749938965,
|
| 471 |
+
"step": 8960
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.4256616322571706,
|
| 475 |
+
"grad_norm": 0.45747169852256775,
|
| 476 |
+
"learning_rate": 0.0006369239847984517,
|
| 477 |
+
"loss": 6.855817794799805,
|
| 478 |
+
"step": 9216
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.4256616322571706,
|
| 482 |
+
"eval_ce_loss": 6.643397864685755,
|
| 483 |
+
"eval_cos_loss": 0.1082146630099375,
|
| 484 |
+
"eval_loss": 6.860041695642689,
|
| 485 |
+
"eval_mse_loss": 0.21664383299818865,
|
| 486 |
+
"flow/cos_sim": 0.8917853384529619,
|
| 487 |
+
"flow/improvement_ratio": 0.990278246740228,
|
| 488 |
+
"flow/mag_ratio_mean": 0.8969191525080432,
|
| 489 |
+
"flow/mag_ratio_std": 0.17551493791004294,
|
| 490 |
+
"step": 9216
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 0.4256616322571706,
|
| 494 |
+
"eval_ce_loss": 6.643397864685755,
|
| 495 |
+
"eval_cos_loss": 0.1082146630099375,
|
| 496 |
+
"eval_loss": 6.860041695642689,
|
| 497 |
+
"eval_mse_loss": 0.21664383299818865,
|
| 498 |
+
"eval_runtime": 64.562,
|
| 499 |
+
"eval_samples_per_second": 433.583,
|
| 500 |
+
"eval_steps_per_second": 6.784,
|
| 501 |
+
"flow/cos_sim": 0.8917853384529619,
|
| 502 |
+
"flow/improvement_ratio": 0.990278246740228,
|
| 503 |
+
"flow/mag_ratio_mean": 0.8969191525080432,
|
| 504 |
+
"flow/mag_ratio_std": 0.17551493791004294,
|
| 505 |
+
"step": 9216
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 0.4374855664865364,
|
| 509 |
+
"grad_norm": 0.529183030128479,
|
| 510 |
+
"learning_rate": 0.0006185336574197479,
|
| 511 |
+
"loss": 6.849204063415527,
|
| 512 |
+
"step": 9472
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.44930950071590225,
|
| 516 |
+
"grad_norm": 0.2831822633743286,
|
| 517 |
+
"learning_rate": 0.0005999717763379407,
|
| 518 |
+
"loss": 6.8596038818359375,
|
| 519 |
+
"step": 9728
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.4611334349452681,
|
| 523 |
+
"grad_norm": 1.0217152833938599,
|
| 524 |
+
"learning_rate": 0.0005812652061542363,
|
| 525 |
+
"loss": 6.847334861755371,
|
| 526 |
+
"step": 9984
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 0.472957369174634,
|
| 530 |
+
"grad_norm": 0.44654422998428345,
|
| 531 |
+
"learning_rate": 0.0005624410208783071,
|
| 532 |
+
"loss": 6.846036911010742,
|
| 533 |
+
"step": 10240
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 0.472957369174634,
|
| 537 |
+
"eval_ce_loss": 6.641612795389951,
|
| 538 |
+
"eval_cos_loss": 0.10159792694939326,
|
| 539 |
+
"eval_loss": 6.845016151802725,
|
| 540 |
+
"eval_mse_loss": 0.20340334783950353,
|
| 541 |
+
"flow/cos_sim": 0.8984020736119519,
|
| 542 |
+
"flow/improvement_ratio": 0.9897379034186062,
|
| 543 |
+
"flow/mag_ratio_mean": 0.9109135144921743,
|
| 544 |
+
"flow/mag_ratio_std": 0.16760862588065945,
|
| 545 |
+
"step": 10240
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 0.472957369174634,
|
| 549 |
+
"eval_ce_loss": 6.641612795389951,
|
| 550 |
+
"eval_cos_loss": 0.10159792694939326,
|
| 551 |
+
"eval_loss": 6.845016151802725,
|
| 552 |
+
"eval_mse_loss": 0.20340334783950353,
|
| 553 |
+
"eval_runtime": 63.9023,
|
| 554 |
+
"eval_samples_per_second": 438.059,
|
| 555 |
+
"eval_steps_per_second": 6.854,
|
| 556 |
+
"flow/cos_sim": 0.8984020736119519,
|
| 557 |
+
"flow/improvement_ratio": 0.9897379034186062,
|
| 558 |
+
"flow/mag_ratio_mean": 0.9109135144921743,
|
| 559 |
+
"flow/mag_ratio_std": 0.16760862588065945,
|
| 560 |
+
"step": 10240
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.48478130340399983,
|
| 564 |
+
"grad_norm": 0.31733208894729614,
|
| 565 |
+
"learning_rate": 0.0005435264647440881,
|
| 566 |
+
"loss": 6.841334342956543,
|
| 567 |
+
"step": 10496
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 0.49660523763336567,
|
| 571 |
+
"grad_norm": 0.7933322191238403,
|
| 572 |
+
"learning_rate": 0.000524548912779213,
|
| 573 |
+
"loss": 6.833290100097656,
|
| 574 |
+
"step": 10752
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 0.5084291718627315,
|
| 578 |
+
"grad_norm": 0.273179829120636,
|
| 579 |
+
"learning_rate": 0.0005055358311851499,
|
| 580 |
+
"loss": 6.844526290893555,
|
| 581 |
+
"step": 11008
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 0.5202531060920974,
|
| 585 |
+
"grad_norm": 0.3748645782470703,
|
| 586 |
+
"learning_rate": 0.0004865147375853812,
|
| 587 |
+
"loss": 6.834465026855469,
|
| 588 |
+
"step": 11264
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 0.5202531060920974,
|
| 592 |
+
"eval_ce_loss": 6.6411590695925495,
|
| 593 |
+
"eval_cos_loss": 0.09674425246236532,
|
| 594 |
+
"eval_loss": 6.83495632480813,
|
| 595 |
+
"eval_mse_loss": 0.19379725269803175,
|
| 596 |
+
"flow/cos_sim": 0.9032557489665132,
|
| 597 |
+
"flow/improvement_ratio": 0.9913779068483065,
|
| 598 |
+
"flow/mag_ratio_mean": 0.916442300631031,
|
| 599 |
+
"flow/mag_ratio_std": 0.16842460159433487,
|
| 600 |
+
"step": 11264
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 0.5202531060920974,
|
| 604 |
+
"eval_ce_loss": 6.6411590695925495,
|
| 605 |
+
"eval_cos_loss": 0.09674425246236532,
|
| 606 |
+
"eval_loss": 6.83495632480813,
|
| 607 |
+
"eval_mse_loss": 0.19379725269803175,
|
| 608 |
+
"eval_runtime": 63.6545,
|
| 609 |
+
"eval_samples_per_second": 439.765,
|
| 610 |
+
"eval_steps_per_second": 6.881,
|
| 611 |
+
"flow/cos_sim": 0.9032557489665132,
|
| 612 |
+
"flow/improvement_ratio": 0.9913779068483065,
|
| 613 |
+
"flow/mag_ratio_mean": 0.916442300631031,
|
| 614 |
+
"flow/mag_ratio_std": 0.16842460159433487,
|
| 615 |
+
"step": 11264
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.5320770403214632,
|
| 619 |
+
"grad_norm": 0.3057026267051697,
|
| 620 |
+
"learning_rate": 0.0004675131611991607,
|
| 621 |
+
"loss": 6.823686599731445,
|
| 622 |
+
"step": 11520
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 0.5439009745508291,
|
| 626 |
+
"grad_norm": 0.6613107919692993,
|
| 627 |
+
"learning_rate": 0.0004485586029984899,
|
| 628 |
+
"loss": 6.826196670532227,
|
| 629 |
+
"step": 11776
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 0.5557249087801949,
|
| 633 |
+
"grad_norm": 0.7032549381256104,
|
| 634 |
+
"learning_rate": 0.00042967849590597266,
|
| 635 |
+
"loss": 6.824607849121094,
|
| 636 |
+
"step": 12032
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 0.5675488430095608,
|
| 640 |
+
"grad_norm": 0.3788175880908966,
|
| 641 |
+
"learning_rate": 0.0004109001650911621,
|
| 642 |
+
"loss": 6.826193332672119,
|
| 643 |
+
"step": 12288
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 0.5675488430095608,
|
| 647 |
+
"eval_ce_loss": 6.639694696147692,
|
| 648 |
+
"eval_cos_loss": 0.0918180431371138,
|
| 649 |
+
"eval_loss": 6.82323066937869,
|
| 650 |
+
"eval_mse_loss": 0.18353597942280442,
|
| 651 |
+
"flow/cos_sim": 0.9081819589551725,
|
| 652 |
+
"flow/improvement_ratio": 0.9904580299984919,
|
| 653 |
+
"flow/mag_ratio_mean": 0.9139516707424704,
|
| 654 |
+
"flow/mag_ratio_std": 0.15966295250202422,
|
| 655 |
+
"step": 12288
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 0.5675488430095608,
|
| 659 |
+
"eval_ce_loss": 6.639694696147692,
|
| 660 |
+
"eval_cos_loss": 0.0918180431371138,
|
| 661 |
+
"eval_loss": 6.82323066937869,
|
| 662 |
+
"eval_mse_loss": 0.18353597942280442,
|
| 663 |
+
"eval_runtime": 63.9479,
|
| 664 |
+
"eval_samples_per_second": 437.747,
|
| 665 |
+
"eval_steps_per_second": 6.849,
|
| 666 |
+
"flow/cos_sim": 0.9081819589551725,
|
| 667 |
+
"flow/improvement_ratio": 0.9904580299984919,
|
| 668 |
+
"flow/mag_ratio_mean": 0.9139516707424704,
|
| 669 |
+
"flow/mag_ratio_std": 0.15966295250202422,
|
| 670 |
+
"step": 12288
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 0.5793727772389267,
|
| 674 |
+
"grad_norm": 0.3380263149738312,
|
| 675 |
+
"learning_rate": 0.0003922507884228551,
|
| 676 |
+
"loss": 6.8268632888793945,
|
| 677 |
+
"step": 12544
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 0.5911967114682924,
|
| 681 |
+
"grad_norm": 0.3507392704486847,
|
| 682 |
+
"learning_rate": 0.00037375735713457723,
|
| 683 |
+
"loss": 6.814322471618652,
|
| 684 |
+
"step": 12800
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 0.6030206456976583,
|
| 688 |
+
"grad_norm": 0.34482619166374207,
|
| 689 |
+
"learning_rate": 0.00035544663676018276,
|
| 690 |
+
"loss": 6.817571640014648,
|
| 691 |
+
"step": 13056
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 0.6148445799270241,
|
| 695 |
+
"grad_norm": 0.3506380617618561,
|
| 696 |
+
"learning_rate": 0.00033734512839611255,
|
| 697 |
+
"loss": 6.815121173858643,
|
| 698 |
+
"step": 13312
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.6148445799270241,
|
| 702 |
+
"eval_ce_loss": 6.639219018422305,
|
| 703 |
+
"eval_cos_loss": 0.08840062753318652,
|
| 704 |
+
"eval_loss": 6.816446893291387,
|
| 705 |
+
"eval_mse_loss": 0.17722788187738966,
|
| 706 |
+
"flow/cos_sim": 0.9115993724838239,
|
| 707 |
+
"flow/improvement_ratio": 0.9913340507576999,
|
| 708 |
+
"flow/mag_ratio_mean": 0.9105819005127911,
|
| 709 |
+
"flow/mag_ratio_std": 0.15417975639779818,
|
| 710 |
+
"step": 13312
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.6148445799270241,
|
| 714 |
+
"eval_ce_loss": 6.639219018422305,
|
| 715 |
+
"eval_cos_loss": 0.08840062753318652,
|
| 716 |
+
"eval_loss": 6.816446893291387,
|
| 717 |
+
"eval_mse_loss": 0.17722788187738966,
|
| 718 |
+
"eval_runtime": 64.1691,
|
| 719 |
+
"eval_samples_per_second": 436.238,
|
| 720 |
+
"eval_steps_per_second": 6.826,
|
| 721 |
+
"flow/cos_sim": 0.9115993724838239,
|
| 722 |
+
"flow/improvement_ratio": 0.9913340507576999,
|
| 723 |
+
"flow/mag_ratio_mean": 0.9105819005127911,
|
| 724 |
+
"flow/mag_ratio_std": 0.15417975639779818,
|
| 725 |
+
"step": 13312
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 0.62666851415639,
|
| 729 |
+
"grad_norm": 0.39763471484184265,
|
| 730 |
+
"learning_rate": 0.0003194790303463687,
|
| 731 |
+
"loss": 6.808745861053467,
|
| 732 |
+
"step": 13568
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 0.6384924483857558,
|
| 736 |
+
"grad_norm": 0.2740715444087982,
|
| 737 |
+
"learning_rate": 0.00030187420020572406,
|
| 738 |
+
"loss": 6.8159685134887695,
|
| 739 |
+
"step": 13824
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 0.6503163826151217,
|
| 743 |
+
"grad_norm": 0.36115965247154236,
|
| 744 |
+
"learning_rate": 0.00028455611743603626,
|
| 745 |
+
"loss": 6.8075947761535645,
|
| 746 |
+
"step": 14080
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.6621403168444876,
|
| 750 |
+
"grad_norm": 0.33763498067855835,
|
| 751 |
+
"learning_rate": 0.0002675498464898373,
|
| 752 |
+
"loss": 6.80622673034668,
|
| 753 |
+
"step": 14336
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 0.6621403168444876,
|
| 757 |
+
"eval_ce_loss": 6.63860485216254,
|
| 758 |
+
"eval_cos_loss": 0.08730801206813556,
|
| 759 |
+
"eval_loss": 6.812814682041673,
|
| 760 |
+
"eval_mse_loss": 0.17420982753169045,
|
| 761 |
+
"flow/cos_sim": 0.9126919905344645,
|
| 762 |
+
"flow/improvement_ratio": 0.9905721886789418,
|
| 763 |
+
"flow/mag_ratio_mean": 0.920888199517716,
|
| 764 |
+
"flow/mag_ratio_std": 0.1555312630044271,
|
| 765 |
+
"step": 14336
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.6621403168444876,
|
| 769 |
+
"eval_ce_loss": 6.63860485216254,
|
| 770 |
+
"eval_cos_loss": 0.08730801206813556,
|
| 771 |
+
"eval_loss": 6.812814682041673,
|
| 772 |
+
"eval_mse_loss": 0.17420982753169045,
|
| 773 |
+
"eval_runtime": 64.0106,
|
| 774 |
+
"eval_samples_per_second": 437.318,
|
| 775 |
+
"eval_steps_per_second": 6.843,
|
| 776 |
+
"flow/cos_sim": 0.9126919905344645,
|
| 777 |
+
"flow/improvement_ratio": 0.9905721886789418,
|
| 778 |
+
"flow/mag_ratio_mean": 0.920888199517716,
|
| 779 |
+
"flow/mag_ratio_std": 0.1555312630044271,
|
| 780 |
+
"step": 14336
|
| 781 |
+
}
|
| 782 |
+
],
|
| 783 |
+
"logging_steps": 256,
|
| 784 |
+
"max_steps": 21651,
|
| 785 |
+
"num_input_tokens_seen": 0,
|
| 786 |
+
"num_train_epochs": 1,
|
| 787 |
+
"save_steps": 1024,
|
| 788 |
+
"stateful_callbacks": {
|
| 789 |
+
"TrainerControl": {
|
| 790 |
+
"args": {
|
| 791 |
+
"should_epoch_stop": false,
|
| 792 |
+
"should_evaluate": false,
|
| 793 |
+
"should_log": false,
|
| 794 |
+
"should_save": true,
|
| 795 |
+
"should_training_stop": false
|
| 796 |
+
},
|
| 797 |
+
"attributes": {}
|
| 798 |
+
}
|
| 799 |
+
},
|
| 800 |
+
"total_flos": 0.0,
|
| 801 |
+
"train_batch_size": 64,
|
| 802 |
+
"trial_name": null,
|
| 803 |
+
"trial_params": null
|
| 804 |
+
}
|
checkpoints-v4.3+/checkpoint-14336/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ee9cb7ebeac20164f12c8a5bf77bbe01722a7871151494aba288a393bf7de081
|
| 3 |
+
size 5137
|