Upload folder using huggingface_hub
Browse files- .gitattributes +1 -0
- checkpoints-v5.6/checkpoint-14336/ema.safetensors +3 -0
- checkpoints-v5.6/checkpoint-14336/eval_state.json +3 -0
- checkpoints-v5.6/checkpoint-14336/model.safetensors +3 -0
- checkpoints-v5.6/checkpoint-14336/optimizer.pt +3 -0
- checkpoints-v5.6/checkpoint-14336/rng_state.pth +3 -0
- checkpoints-v5.6/checkpoint-14336/scaler.pt +3 -0
- checkpoints-v5.6/checkpoint-14336/scheduler.pt +3 -0
- checkpoints-v5.6/checkpoint-14336/trainer_state.json +566 -0
- checkpoints-v5.6/checkpoint-14336/training_args.bin +3 -0
.gitattributes
CHANGED
|
@@ -43,3 +43,4 @@ checkpoints-v5.5/checkpoint-16384/eval_state.json filter=lfs diff=lfs merge=lfs
|
|
| 43 |
checkpoints-v4.4+/checkpoint-7168/eval_state.json filter=lfs diff=lfs merge=lfs -text
|
| 44 |
checkpoints-v5.5/checkpoint-24576/eval_state.json filter=lfs diff=lfs merge=lfs -text
|
| 45 |
checkpoints-v5.6/checkpoint-4096/eval_state.json filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 43 |
checkpoints-v4.4+/checkpoint-7168/eval_state.json filter=lfs diff=lfs merge=lfs -text
|
| 44 |
checkpoints-v5.5/checkpoint-24576/eval_state.json filter=lfs diff=lfs merge=lfs -text
|
| 45 |
checkpoints-v5.6/checkpoint-4096/eval_state.json filter=lfs diff=lfs merge=lfs -text
|
| 46 |
+
checkpoints-v5.6/checkpoint-14336/eval_state.json filter=lfs diff=lfs merge=lfs -text
|
checkpoints-v5.6/checkpoint-14336/ema.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f06b4cb6ab59aa4e5cfc27af41908925b25c9a1919b4fffb002fccd47fb83dd5
|
| 3 |
+
size 54599592
|
checkpoints-v5.6/checkpoint-14336/eval_state.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:faec1ef62aef8e8e3974a59295f05c1d11d0e69ead203ad058247a8ada06311a
|
| 3 |
+
size 58408752
|
checkpoints-v5.6/checkpoint-14336/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d658416497177d36a2a33f05ef6c531a5d65993c70c4a9d3138885f6e36463f9
|
| 3 |
+
size 54599624
|
checkpoints-v5.6/checkpoint-14336/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8f9fc4bfb123840f05ffc8b78cccd5d95190ec220ea328196eefeced9dbf54af
|
| 3 |
+
size 76550347
|
checkpoints-v5.6/checkpoint-14336/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8c84b0b910e410334974b64277a31060a45a30a6af56f606c7714596a8a3a85d
|
| 3 |
+
size 14645
|
checkpoints-v5.6/checkpoint-14336/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:79701bbdc0c49714e0e085b5c67881fa167d2bf0d010e2195ebdf057057d72dd
|
| 3 |
+
size 1383
|
checkpoints-v5.6/checkpoint-14336/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3fea187e43031df6b1b3d8be1593743f072013fecb64ec1a061b4e69f4994c94
|
| 3 |
+
size 1465
|
checkpoints-v5.6/checkpoint-14336/trainer_state.json
ADDED
|
@@ -0,0 +1,566 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.6178378261900145,
|
| 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": 0.04413127329928675,
|
| 14 |
+
"grad_norm": 2.024094581604004,
|
| 15 |
+
"learning_rate": 9.990234375e-05,
|
| 16 |
+
"loss": 13.011528968811035,
|
| 17 |
+
"step": 1024
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.04413127329928675,
|
| 21 |
+
"eval_bleu": 0.13973491511585304,
|
| 22 |
+
"eval_ce_clean_loss": 2.321801658886582,
|
| 23 |
+
"eval_ce_pred_loss": 5.398845981687371,
|
| 24 |
+
"eval_flow_cos_loss": 0.4359625861970092,
|
| 25 |
+
"eval_flow_mse_loss": 1.261753735003441,
|
| 26 |
+
"eval_loss": 7.471738189014037,
|
| 27 |
+
"flow/cos_sim": 0.5640374244784495,
|
| 28 |
+
"flow/improvement_ratio": 0.9125845870737836,
|
| 29 |
+
"flow/mag_ratio_mean": 0.005239322681281803,
|
| 30 |
+
"flow/mag_ratio_std": 0.0019533936416349018,
|
| 31 |
+
"step": 1024
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.04413127329928675,
|
| 35 |
+
"eval_bleu": 0.13973491511585304,
|
| 36 |
+
"eval_ce_clean_loss": 2.321801658886582,
|
| 37 |
+
"eval_ce_pred_loss": 5.398845981687371,
|
| 38 |
+
"eval_flow_cos_loss": 0.4359625861970092,
|
| 39 |
+
"eval_flow_mse_loss": 1.261753735003441,
|
| 40 |
+
"eval_loss": 7.471738189014037,
|
| 41 |
+
"eval_runtime": 200.3486,
|
| 42 |
+
"eval_samples_per_second": 149.739,
|
| 43 |
+
"eval_steps_per_second": 2.341,
|
| 44 |
+
"flow/cos_sim": 0.5640374244784495,
|
| 45 |
+
"flow/improvement_ratio": 0.9125845870737836,
|
| 46 |
+
"flow/mag_ratio_mean": 0.005239322681281803,
|
| 47 |
+
"flow/mag_ratio_std": 0.0019533936416349018,
|
| 48 |
+
"step": 1024
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 0.0882625465985735,
|
| 52 |
+
"grad_norm": 2.3836026191711426,
|
| 53 |
+
"learning_rate": 9.9476028157316e-05,
|
| 54 |
+
"loss": 5.076037406921387,
|
| 55 |
+
"step": 2048
|
| 56 |
+
},
|
| 57 |
+
{
|
| 58 |
+
"epoch": 0.0882625465985735,
|
| 59 |
+
"eval_bleu": 0.334789938692836,
|
| 60 |
+
"eval_ce_clean_loss": 0.34960190961355847,
|
| 61 |
+
"eval_ce_pred_loss": 3.634908381301457,
|
| 62 |
+
"eval_flow_cos_loss": 0.4467242890075326,
|
| 63 |
+
"eval_flow_mse_loss": 1.0751057558222366,
|
| 64 |
+
"eval_loss": 4.080824585611632,
|
| 65 |
+
"flow/cos_sim": 0.5532757186177951,
|
| 66 |
+
"flow/improvement_ratio": 0.988054730363492,
|
| 67 |
+
"flow/mag_ratio_mean": 0.4422806202094438,
|
| 68 |
+
"flow/mag_ratio_std": 0.11935382749416681,
|
| 69 |
+
"step": 2048
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.0882625465985735,
|
| 73 |
+
"eval_bleu": 0.334789938692836,
|
| 74 |
+
"eval_ce_clean_loss": 0.34960190961355847,
|
| 75 |
+
"eval_ce_pred_loss": 3.634908381301457,
|
| 76 |
+
"eval_flow_cos_loss": 0.4467242890075326,
|
| 77 |
+
"eval_flow_mse_loss": 1.0751057558222366,
|
| 78 |
+
"eval_loss": 4.080824585611632,
|
| 79 |
+
"eval_runtime": 198.1742,
|
| 80 |
+
"eval_samples_per_second": 151.382,
|
| 81 |
+
"eval_steps_per_second": 2.367,
|
| 82 |
+
"flow/cos_sim": 0.5532757186177951,
|
| 83 |
+
"flow/improvement_ratio": 0.988054730363492,
|
| 84 |
+
"flow/mag_ratio_mean": 0.4422806202094438,
|
| 85 |
+
"flow/mag_ratio_std": 0.11935382749416681,
|
| 86 |
+
"step": 2048
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.13239381989786023,
|
| 90 |
+
"grad_norm": 0.8187930583953857,
|
| 91 |
+
"learning_rate": 9.791307026072513e-05,
|
| 92 |
+
"loss": 3.823514938354492,
|
| 93 |
+
"step": 3072
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.13239381989786023,
|
| 97 |
+
"eval_bleu": 0.372073011692056,
|
| 98 |
+
"eval_ce_clean_loss": 0.13431392060414052,
|
| 99 |
+
"eval_ce_pred_loss": 3.167514971832731,
|
| 100 |
+
"eval_flow_cos_loss": 0.3423818138870857,
|
| 101 |
+
"eval_flow_mse_loss": 1.1092233459578393,
|
| 102 |
+
"eval_loss": 3.5463931677438048,
|
| 103 |
+
"flow/cos_sim": 0.6576181853503815,
|
| 104 |
+
"flow/improvement_ratio": 0.9944826788993787,
|
| 105 |
+
"flow/mag_ratio_mean": 0.6025184103166625,
|
| 106 |
+
"flow/mag_ratio_std": 0.11320645424094536,
|
| 107 |
+
"step": 3072
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.13239381989786023,
|
| 111 |
+
"eval_bleu": 0.372073011692056,
|
| 112 |
+
"eval_ce_clean_loss": 0.13431392060414052,
|
| 113 |
+
"eval_ce_pred_loss": 3.167514971832731,
|
| 114 |
+
"eval_flow_cos_loss": 0.3423818138870857,
|
| 115 |
+
"eval_flow_mse_loss": 1.1092233459578393,
|
| 116 |
+
"eval_loss": 3.5463931677438048,
|
| 117 |
+
"eval_runtime": 199.1274,
|
| 118 |
+
"eval_samples_per_second": 150.657,
|
| 119 |
+
"eval_steps_per_second": 2.355,
|
| 120 |
+
"flow/cos_sim": 0.6576181853503815,
|
| 121 |
+
"flow/improvement_ratio": 0.9944826788993787,
|
| 122 |
+
"flow/mag_ratio_mean": 0.6025184103166625,
|
| 123 |
+
"flow/mag_ratio_std": 0.11320645424094536,
|
| 124 |
+
"step": 3072
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 0.176525093197147,
|
| 128 |
+
"grad_norm": 1.596978783607483,
|
| 129 |
+
"learning_rate": 9.53439476074686e-05,
|
| 130 |
+
"loss": 3.4521546363830566,
|
| 131 |
+
"step": 4096
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 0.176525093197147,
|
| 135 |
+
"eval_bleu": 0.3922768378761595,
|
| 136 |
+
"eval_ce_clean_loss": 0.06992586965023327,
|
| 137 |
+
"eval_ce_pred_loss": 2.9707689447951977,
|
| 138 |
+
"eval_flow_cos_loss": 0.18835138766241988,
|
| 139 |
+
"eval_flow_mse_loss": 1.08429173365839,
|
| 140 |
+
"eval_loss": 3.2808436879725345,
|
| 141 |
+
"flow/cos_sim": 0.8116486336884976,
|
| 142 |
+
"flow/improvement_ratio": 0.9945465602091889,
|
| 143 |
+
"flow/mag_ratio_mean": 0.7910831776509153,
|
| 144 |
+
"flow/mag_ratio_std": 0.09716444489544135,
|
| 145 |
+
"step": 4096
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 0.176525093197147,
|
| 149 |
+
"eval_bleu": 0.3922768378761595,
|
| 150 |
+
"eval_ce_clean_loss": 0.06992586965023327,
|
| 151 |
+
"eval_ce_pred_loss": 2.9707689447951977,
|
| 152 |
+
"eval_flow_cos_loss": 0.18835138766241988,
|
| 153 |
+
"eval_flow_mse_loss": 1.08429173365839,
|
| 154 |
+
"eval_loss": 3.2808436879725345,
|
| 155 |
+
"eval_runtime": 199.2079,
|
| 156 |
+
"eval_samples_per_second": 150.596,
|
| 157 |
+
"eval_steps_per_second": 2.354,
|
| 158 |
+
"flow/cos_sim": 0.8116486336884976,
|
| 159 |
+
"flow/improvement_ratio": 0.9945465602091889,
|
| 160 |
+
"flow/mag_ratio_mean": 0.7910831776509153,
|
| 161 |
+
"flow/mag_ratio_std": 0.09716444489544135,
|
| 162 |
+
"step": 4096
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 0.22065636649643372,
|
| 166 |
+
"grad_norm": 0.8742256760597229,
|
| 167 |
+
"learning_rate": 9.18264920723673e-05,
|
| 168 |
+
"loss": 3.2747244834899902,
|
| 169 |
+
"step": 5120
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 0.22065636649643372,
|
| 173 |
+
"eval_bleu": 0.39936848850149986,
|
| 174 |
+
"eval_ce_clean_loss": 0.04289639472310096,
|
| 175 |
+
"eval_ce_pred_loss": 2.848231249780797,
|
| 176 |
+
"eval_flow_cos_loss": 0.13245032197122636,
|
| 177 |
+
"eval_flow_mse_loss": 1.07859928343596,
|
| 178 |
+
"eval_loss": 3.148370101253615,
|
| 179 |
+
"flow/cos_sim": 0.867549685908279,
|
| 180 |
+
"flow/improvement_ratio": 0.9947950637925154,
|
| 181 |
+
"flow/mag_ratio_mean": 0.8595841084716163,
|
| 182 |
+
"flow/mag_ratio_std": 0.08974277026363528,
|
| 183 |
+
"step": 5120
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.22065636649643372,
|
| 187 |
+
"eval_bleu": 0.39936848850149986,
|
| 188 |
+
"eval_ce_clean_loss": 0.04289639472310096,
|
| 189 |
+
"eval_ce_pred_loss": 2.848231249780797,
|
| 190 |
+
"eval_flow_cos_loss": 0.13245032197122636,
|
| 191 |
+
"eval_flow_mse_loss": 1.07859928343596,
|
| 192 |
+
"eval_loss": 3.148370101253615,
|
| 193 |
+
"eval_runtime": 207.1188,
|
| 194 |
+
"eval_samples_per_second": 144.844,
|
| 195 |
+
"eval_steps_per_second": 2.264,
|
| 196 |
+
"flow/cos_sim": 0.867549685908279,
|
| 197 |
+
"flow/improvement_ratio": 0.9947950637925154,
|
| 198 |
+
"flow/mag_ratio_mean": 0.8595841084716163,
|
| 199 |
+
"flow/mag_ratio_std": 0.08974277026363528,
|
| 200 |
+
"step": 5120
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.26478763979572045,
|
| 204 |
+
"grad_norm": 1.2167394161224365,
|
| 205 |
+
"learning_rate": 8.742770483354739e-05,
|
| 206 |
+
"loss": 3.171807050704956,
|
| 207 |
+
"step": 6144
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.26478763979572045,
|
| 211 |
+
"eval_bleu": 0.42114677949263607,
|
| 212 |
+
"eval_ce_clean_loss": 0.02962892580983926,
|
| 213 |
+
"eval_ce_pred_loss": 2.721838645081022,
|
| 214 |
+
"eval_flow_cos_loss": 0.11157757617326687,
|
| 215 |
+
"eval_flow_mse_loss": 1.0874363405109724,
|
| 216 |
+
"eval_loss": 3.0502466992782886,
|
| 217 |
+
"flow/cos_sim": 0.8884224528188644,
|
| 218 |
+
"flow/improvement_ratio": 0.994537014061454,
|
| 219 |
+
"flow/mag_ratio_mean": 0.882329723601148,
|
| 220 |
+
"flow/mag_ratio_std": 0.08614625519653882,
|
| 221 |
+
"step": 6144
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.26478763979572045,
|
| 225 |
+
"eval_bleu": 0.42114677949263607,
|
| 226 |
+
"eval_ce_clean_loss": 0.02962892580983926,
|
| 227 |
+
"eval_ce_pred_loss": 2.721838645081022,
|
| 228 |
+
"eval_flow_cos_loss": 0.11157757617326687,
|
| 229 |
+
"eval_flow_mse_loss": 1.0874363405109724,
|
| 230 |
+
"eval_loss": 3.0502466992782886,
|
| 231 |
+
"eval_runtime": 201.5414,
|
| 232 |
+
"eval_samples_per_second": 148.853,
|
| 233 |
+
"eval_steps_per_second": 2.327,
|
| 234 |
+
"flow/cos_sim": 0.8884224528188644,
|
| 235 |
+
"flow/improvement_ratio": 0.994537014061454,
|
| 236 |
+
"flow/mag_ratio_mean": 0.882329723601148,
|
| 237 |
+
"flow/mag_ratio_std": 0.08614625519653882,
|
| 238 |
+
"step": 6144
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 0.30891891309500724,
|
| 242 |
+
"grad_norm": 1.7619376182556152,
|
| 243 |
+
"learning_rate": 8.22483558761947e-05,
|
| 244 |
+
"loss": 3.073448896408081,
|
| 245 |
+
"step": 7168
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 0.30891891309500724,
|
| 249 |
+
"eval_bleu": 0.4370876849406368,
|
| 250 |
+
"eval_ce_clean_loss": 0.02222317859911715,
|
| 251 |
+
"eval_ce_pred_loss": 2.6402354730980226,
|
| 252 |
+
"eval_flow_cos_loss": 0.10016548941765767,
|
| 253 |
+
"eval_flow_mse_loss": 1.0842415976371846,
|
| 254 |
+
"eval_loss": 2.979670949582098,
|
| 255 |
+
"flow/cos_sim": 0.8998345271356578,
|
| 256 |
+
"flow/improvement_ratio": 0.9956297869367132,
|
| 257 |
+
"flow/mag_ratio_mean": 0.8991066802030941,
|
| 258 |
+
"flow/mag_ratio_std": 0.08455248013424721,
|
| 259 |
+
"step": 7168
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 0.30891891309500724,
|
| 263 |
+
"eval_bleu": 0.4370876849406368,
|
| 264 |
+
"eval_ce_clean_loss": 0.02222317859911715,
|
| 265 |
+
"eval_ce_pred_loss": 2.6402354730980226,
|
| 266 |
+
"eval_flow_cos_loss": 0.10016548941765767,
|
| 267 |
+
"eval_flow_mse_loss": 1.0842415976371846,
|
| 268 |
+
"eval_loss": 2.979670949582098,
|
| 269 |
+
"eval_runtime": 203.5883,
|
| 270 |
+
"eval_samples_per_second": 147.356,
|
| 271 |
+
"eval_steps_per_second": 2.304,
|
| 272 |
+
"flow/cos_sim": 0.8998345271356578,
|
| 273 |
+
"flow/improvement_ratio": 0.9956297869367132,
|
| 274 |
+
"flow/mag_ratio_mean": 0.8991066802030941,
|
| 275 |
+
"flow/mag_ratio_std": 0.08455248013424721,
|
| 276 |
+
"step": 7168
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.353050186394294,
|
| 280 |
+
"grad_norm": 1.8936859369277954,
|
| 281 |
+
"learning_rate": 7.639311770076283e-05,
|
| 282 |
+
"loss": 3.0209758281707764,
|
| 283 |
+
"step": 8192
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.353050186394294,
|
| 287 |
+
"eval_bleu": 0.4462124473905187,
|
| 288 |
+
"eval_ce_clean_loss": 0.017575070791160947,
|
| 289 |
+
"eval_ce_pred_loss": 2.586446302277701,
|
| 290 |
+
"eval_flow_cos_loss": 0.09220715248381406,
|
| 291 |
+
"eval_flow_mse_loss": 1.0823260478373529,
|
| 292 |
+
"eval_loss": 2.933465297288224,
|
| 293 |
+
"flow/cos_sim": 0.9077928694072308,
|
| 294 |
+
"flow/improvement_ratio": 0.9947618440524347,
|
| 295 |
+
"flow/mag_ratio_mean": 0.8988318181495423,
|
| 296 |
+
"flow/mag_ratio_std": 0.08265599157256104,
|
| 297 |
+
"step": 8192
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.353050186394294,
|
| 301 |
+
"eval_bleu": 0.4462124473905187,
|
| 302 |
+
"eval_ce_clean_loss": 0.017575070791160947,
|
| 303 |
+
"eval_ce_pred_loss": 2.586446302277701,
|
| 304 |
+
"eval_flow_cos_loss": 0.09220715248381406,
|
| 305 |
+
"eval_flow_mse_loss": 1.0823260478373529,
|
| 306 |
+
"eval_loss": 2.933465297288224,
|
| 307 |
+
"eval_runtime": 205.9836,
|
| 308 |
+
"eval_samples_per_second": 145.643,
|
| 309 |
+
"eval_steps_per_second": 2.277,
|
| 310 |
+
"flow/cos_sim": 0.9077928694072308,
|
| 311 |
+
"flow/improvement_ratio": 0.9947618440524347,
|
| 312 |
+
"flow/mag_ratio_mean": 0.8988318181495423,
|
| 313 |
+
"flow/mag_ratio_std": 0.08265599157256104,
|
| 314 |
+
"step": 8192
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.3971814596935807,
|
| 318 |
+
"grad_norm": 1.6591744422912598,
|
| 319 |
+
"learning_rate": 6.997821756319211e-05,
|
| 320 |
+
"loss": 2.97641921043396,
|
| 321 |
+
"step": 9216
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.3971814596935807,
|
| 325 |
+
"eval_bleu": 0.4548893990902779,
|
| 326 |
+
"eval_ce_clean_loss": 0.014087324782308421,
|
| 327 |
+
"eval_ce_pred_loss": 2.5131919986403575,
|
| 328 |
+
"eval_flow_cos_loss": 0.08636398176585179,
|
| 329 |
+
"eval_flow_mse_loss": 1.073010201901515,
|
| 330 |
+
"eval_loss": 2.867922893211023,
|
| 331 |
+
"flow/cos_sim": 0.9136360249539682,
|
| 332 |
+
"flow/improvement_ratio": 0.9953014714631445,
|
| 333 |
+
"flow/mag_ratio_mean": 0.9083017234100716,
|
| 334 |
+
"flow/mag_ratio_std": 0.08202346263409678,
|
| 335 |
+
"step": 9216
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 0.3971814596935807,
|
| 339 |
+
"eval_bleu": 0.4548893990902779,
|
| 340 |
+
"eval_ce_clean_loss": 0.014087324782308421,
|
| 341 |
+
"eval_ce_pred_loss": 2.5131919986403575,
|
| 342 |
+
"eval_flow_cos_loss": 0.08636398176585179,
|
| 343 |
+
"eval_flow_mse_loss": 1.073010201901515,
|
| 344 |
+
"eval_loss": 2.867922893211023,
|
| 345 |
+
"eval_runtime": 205.454,
|
| 346 |
+
"eval_samples_per_second": 146.018,
|
| 347 |
+
"eval_steps_per_second": 2.283,
|
| 348 |
+
"flow/cos_sim": 0.9136360249539682,
|
| 349 |
+
"flow/improvement_ratio": 0.9953014714631445,
|
| 350 |
+
"flow/mag_ratio_mean": 0.9083017234100716,
|
| 351 |
+
"flow/mag_ratio_std": 0.08202346263409678,
|
| 352 |
+
"step": 9216
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.44131273299286744,
|
| 356 |
+
"grad_norm": 2.6923091411590576,
|
| 357 |
+
"learning_rate": 6.315061173955019e-05,
|
| 358 |
+
"loss": 2.924818515777588,
|
| 359 |
+
"step": 10240
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.44131273299286744,
|
| 363 |
+
"eval_bleu": 0.4700031951689637,
|
| 364 |
+
"eval_ce_clean_loss": 0.011653113022033593,
|
| 365 |
+
"eval_ce_pred_loss": 2.5109867228627967,
|
| 366 |
+
"eval_flow_cos_loss": 0.08097743840296386,
|
| 367 |
+
"eval_flow_mse_loss": 1.0633719702010977,
|
| 368 |
+
"eval_loss": 2.8529601102190485,
|
| 369 |
+
"flow/cos_sim": 0.9190225931627156,
|
| 370 |
+
"flow/improvement_ratio": 0.9958856232893238,
|
| 371 |
+
"flow/mag_ratio_mean": 0.9080375746877463,
|
| 372 |
+
"flow/mag_ratio_std": 0.07999324749337076,
|
| 373 |
+
"step": 10240
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.44131273299286744,
|
| 377 |
+
"eval_bleu": 0.4700031951689637,
|
| 378 |
+
"eval_ce_clean_loss": 0.011653113022033593,
|
| 379 |
+
"eval_ce_pred_loss": 2.5109867228627967,
|
| 380 |
+
"eval_flow_cos_loss": 0.08097743840296386,
|
| 381 |
+
"eval_flow_mse_loss": 1.0633719702010977,
|
| 382 |
+
"eval_loss": 2.8529601102190485,
|
| 383 |
+
"eval_runtime": 203.8367,
|
| 384 |
+
"eval_samples_per_second": 147.177,
|
| 385 |
+
"eval_steps_per_second": 2.301,
|
| 386 |
+
"flow/cos_sim": 0.9190225931627156,
|
| 387 |
+
"flow/improvement_ratio": 0.9958856232893238,
|
| 388 |
+
"flow/mag_ratio_mean": 0.9080375746877463,
|
| 389 |
+
"flow/mag_ratio_std": 0.07999324749337076,
|
| 390 |
+
"step": 10240
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 0.4854440062921542,
|
| 394 |
+
"grad_norm": 3.9323318004608154,
|
| 395 |
+
"learning_rate": 5.604035379537632e-05,
|
| 396 |
+
"loss": 2.895775556564331,
|
| 397 |
+
"step": 11264
|
| 398 |
+
},
|
| 399 |
+
{
|
| 400 |
+
"epoch": 0.4854440062921542,
|
| 401 |
+
"eval_bleu": 0.4627536040081024,
|
| 402 |
+
"eval_ce_clean_loss": 0.009901690839934769,
|
| 403 |
+
"eval_ce_pred_loss": 2.51465770151061,
|
| 404 |
+
"eval_flow_cos_loss": 0.07650925458939091,
|
| 405 |
+
"eval_flow_mse_loss": 1.041963977218945,
|
| 406 |
+
"eval_loss": 2.8312533544833216,
|
| 407 |
+
"flow/cos_sim": 0.9234907618209497,
|
| 408 |
+
"flow/improvement_ratio": 0.9943228665191227,
|
| 409 |
+
"flow/mag_ratio_mean": 0.9196320714981063,
|
| 410 |
+
"flow/mag_ratio_std": 0.07909934378381986,
|
| 411 |
+
"step": 11264
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 0.4854440062921542,
|
| 415 |
+
"eval_bleu": 0.4627536040081024,
|
| 416 |
+
"eval_ce_clean_loss": 0.009901690839934769,
|
| 417 |
+
"eval_ce_pred_loss": 2.51465770151061,
|
| 418 |
+
"eval_flow_cos_loss": 0.07650925458939091,
|
| 419 |
+
"eval_flow_mse_loss": 1.041963977218945,
|
| 420 |
+
"eval_loss": 2.8312533544833216,
|
| 421 |
+
"eval_runtime": 203.4569,
|
| 422 |
+
"eval_samples_per_second": 147.451,
|
| 423 |
+
"eval_steps_per_second": 2.305,
|
| 424 |
+
"flow/cos_sim": 0.9234907618209497,
|
| 425 |
+
"flow/improvement_ratio": 0.9943228665191227,
|
| 426 |
+
"flow/mag_ratio_mean": 0.9196320714981063,
|
| 427 |
+
"flow/mag_ratio_std": 0.07909934378381986,
|
| 428 |
+
"step": 11264
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.5295752795914409,
|
| 432 |
+
"grad_norm": 1.9154846668243408,
|
| 433 |
+
"learning_rate": 4.881032966918056e-05,
|
| 434 |
+
"loss": 2.879368305206299,
|
| 435 |
+
"step": 12288
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.5295752795914409,
|
| 439 |
+
"eval_bleu": 0.4667095049087505,
|
| 440 |
+
"eval_ce_clean_loss": 0.008838278966258838,
|
| 441 |
+
"eval_ce_pred_loss": 2.4798952907895737,
|
| 442 |
+
"eval_flow_cos_loss": 0.07372279336521112,
|
| 443 |
+
"eval_flow_mse_loss": 1.0375378379689606,
|
| 444 |
+
"eval_loss": 2.8007334864724167,
|
| 445 |
+
"flow/cos_sim": 0.9262772229180407,
|
| 446 |
+
"flow/improvement_ratio": 0.9949917554346992,
|
| 447 |
+
"flow/mag_ratio_mean": 0.9212620670098994,
|
| 448 |
+
"flow/mag_ratio_std": 0.07803667050752558,
|
| 449 |
+
"step": 12288
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 0.5295752795914409,
|
| 453 |
+
"eval_bleu": 0.4667095049087505,
|
| 454 |
+
"eval_ce_clean_loss": 0.008838278966258838,
|
| 455 |
+
"eval_ce_pred_loss": 2.4798952907895737,
|
| 456 |
+
"eval_flow_cos_loss": 0.07372279336521112,
|
| 457 |
+
"eval_flow_mse_loss": 1.0375378379689606,
|
| 458 |
+
"eval_loss": 2.8007334864724167,
|
| 459 |
+
"eval_runtime": 202.9744,
|
| 460 |
+
"eval_samples_per_second": 147.802,
|
| 461 |
+
"eval_steps_per_second": 2.311,
|
| 462 |
+
"flow/cos_sim": 0.9262772229180407,
|
| 463 |
+
"flow/improvement_ratio": 0.9949917554346992,
|
| 464 |
+
"flow/mag_ratio_mean": 0.9212620670098994,
|
| 465 |
+
"flow/mag_ratio_std": 0.07803667050752558,
|
| 466 |
+
"step": 12288
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 0.5737065528907277,
|
| 470 |
+
"grad_norm": 2.099470376968384,
|
| 471 |
+
"learning_rate": 4.159825826870804e-05,
|
| 472 |
+
"loss": 2.84110951423645,
|
| 473 |
+
"step": 13312
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 0.5737065528907277,
|
| 477 |
+
"eval_bleu": 0.4697937916499978,
|
| 478 |
+
"eval_ce_clean_loss": 0.008110439863556357,
|
| 479 |
+
"eval_ce_pred_loss": 2.445156754207001,
|
| 480 |
+
"eval_flow_cos_loss": 0.07064609612419662,
|
| 481 |
+
"eval_flow_mse_loss": 1.030803042942527,
|
| 482 |
+
"eval_loss": 2.7681847221053233,
|
| 483 |
+
"flow/cos_sim": 0.9293539301672978,
|
| 484 |
+
"flow/improvement_ratio": 0.9955378606883701,
|
| 485 |
+
"flow/mag_ratio_mean": 0.9253565624578676,
|
| 486 |
+
"flow/mag_ratio_std": 0.07775539313869945,
|
| 487 |
+
"step": 13312
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"epoch": 0.5737065528907277,
|
| 491 |
+
"eval_bleu": 0.4697937916499978,
|
| 492 |
+
"eval_ce_clean_loss": 0.008110439863556357,
|
| 493 |
+
"eval_ce_pred_loss": 2.445156754207001,
|
| 494 |
+
"eval_flow_cos_loss": 0.07064609612419662,
|
| 495 |
+
"eval_flow_mse_loss": 1.030803042942527,
|
| 496 |
+
"eval_loss": 2.7681847221053233,
|
| 497 |
+
"eval_runtime": 202.522,
|
| 498 |
+
"eval_samples_per_second": 148.132,
|
| 499 |
+
"eval_steps_per_second": 2.316,
|
| 500 |
+
"flow/cos_sim": 0.9293539301672978,
|
| 501 |
+
"flow/improvement_ratio": 0.9955378606883701,
|
| 502 |
+
"flow/mag_ratio_mean": 0.9253565624578676,
|
| 503 |
+
"flow/mag_ratio_std": 0.07775539313869945,
|
| 504 |
+
"step": 13312
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 0.6178378261900145,
|
| 508 |
+
"grad_norm": 5.420510292053223,
|
| 509 |
+
"learning_rate": 3.456935793454373e-05,
|
| 510 |
+
"loss": 2.829457998275757,
|
| 511 |
+
"step": 14336
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 0.6178378261900145,
|
| 515 |
+
"eval_bleu": 0.47719369632160064,
|
| 516 |
+
"eval_ce_clean_loss": 0.007385544868102714,
|
| 517 |
+
"eval_ce_pred_loss": 2.418667951880742,
|
| 518 |
+
"eval_flow_cos_loss": 0.06978921110886754,
|
| 519 |
+
"eval_flow_mse_loss": 1.0323392859399954,
|
| 520 |
+
"eval_loss": 2.7502396734538617,
|
| 521 |
+
"flow/cos_sim": 0.9302108099719862,
|
| 522 |
+
"flow/improvement_ratio": 0.9951674874657507,
|
| 523 |
+
"flow/mag_ratio_mean": 0.9268048170533008,
|
| 524 |
+
"flow/mag_ratio_std": 0.07696014680842092,
|
| 525 |
+
"step": 14336
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 0.6178378261900145,
|
| 529 |
+
"eval_bleu": 0.47719369632160064,
|
| 530 |
+
"eval_ce_clean_loss": 0.007385544868102714,
|
| 531 |
+
"eval_ce_pred_loss": 2.418667951880742,
|
| 532 |
+
"eval_flow_cos_loss": 0.06978921110886754,
|
| 533 |
+
"eval_flow_mse_loss": 1.0323392859399954,
|
| 534 |
+
"eval_loss": 2.7502396734538617,
|
| 535 |
+
"eval_runtime": 202.3423,
|
| 536 |
+
"eval_samples_per_second": 148.264,
|
| 537 |
+
"eval_steps_per_second": 2.318,
|
| 538 |
+
"flow/cos_sim": 0.9302108099719862,
|
| 539 |
+
"flow/improvement_ratio": 0.9951674874657507,
|
| 540 |
+
"flow/mag_ratio_mean": 0.9268048170533008,
|
| 541 |
+
"flow/mag_ratio_std": 0.07696014680842092,
|
| 542 |
+
"step": 14336
|
| 543 |
+
}
|
| 544 |
+
],
|
| 545 |
+
"logging_steps": 1024,
|
| 546 |
+
"max_steps": 23204,
|
| 547 |
+
"num_input_tokens_seen": 0,
|
| 548 |
+
"num_train_epochs": 1,
|
| 549 |
+
"save_steps": 1024,
|
| 550 |
+
"stateful_callbacks": {
|
| 551 |
+
"TrainerControl": {
|
| 552 |
+
"args": {
|
| 553 |
+
"should_epoch_stop": false,
|
| 554 |
+
"should_evaluate": false,
|
| 555 |
+
"should_log": false,
|
| 556 |
+
"should_save": true,
|
| 557 |
+
"should_training_stop": false
|
| 558 |
+
},
|
| 559 |
+
"attributes": {}
|
| 560 |
+
}
|
| 561 |
+
},
|
| 562 |
+
"total_flos": 0.0,
|
| 563 |
+
"train_batch_size": 64,
|
| 564 |
+
"trial_name": null,
|
| 565 |
+
"trial_params": null
|
| 566 |
+
}
|
checkpoints-v5.6/checkpoint-14336/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8469bbc81a2ba0be2f5b44007faafd15c75615abe30f4f4e56171816d31caa5b
|
| 3 |
+
size 5137
|