Upload folder using huggingface_hub
Browse files- .gitattributes +1 -0
- checkpoints-v4.5/checkpoint-15360/ema.safetensors +3 -0
- checkpoints-v4.5/checkpoint-15360/eval_state.json +3 -0
- checkpoints-v4.5/checkpoint-15360/model.safetensors +3 -0
- checkpoints-v4.5/checkpoint-15360/optimizer.pt +3 -0
- checkpoints-v4.5/checkpoint-15360/rng_state.pth +3 -0
- checkpoints-v4.5/checkpoint-15360/scaler.pt +3 -0
- checkpoints-v4.5/checkpoint-15360/scheduler.pt +3 -0
- checkpoints-v4.5/checkpoint-15360/trainer_state.json +919 -0
- checkpoints-v4.5/checkpoint-15360/training_args.bin +3 -0
.gitattributes
CHANGED
|
@@ -47,3 +47,4 @@ checkpoints-v5.6/checkpoint-14336/eval_state.json filter=lfs diff=lfs merge=lfs
|
|
| 47 |
checkpoints-v5.7/checkpoint-11264/eval_state.json filter=lfs diff=lfs merge=lfs -text
|
| 48 |
checkpoints-v5.8/checkpoint-10240/eval_state.json filter=lfs diff=lfs merge=lfs -text
|
| 49 |
checkpoints-v4.5/checkpoint-7168/eval_state.json filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 47 |
checkpoints-v5.7/checkpoint-11264/eval_state.json filter=lfs diff=lfs merge=lfs -text
|
| 48 |
checkpoints-v5.8/checkpoint-10240/eval_state.json filter=lfs diff=lfs merge=lfs -text
|
| 49 |
checkpoints-v4.5/checkpoint-7168/eval_state.json filter=lfs diff=lfs merge=lfs -text
|
| 50 |
+
checkpoints-v4.5/checkpoint-15360/eval_state.json filter=lfs diff=lfs merge=lfs -text
|
checkpoints-v4.5/checkpoint-15360/ema.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e9abdc0fe24770d700d9543b9b4cfee7f1aa5e9bb0bda8bb531f39970e0e6712
|
| 3 |
+
size 38935592
|
checkpoints-v4.5/checkpoint-15360/eval_state.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d838ebc3cb2fb06e8e55705aa23b11bc858a6700265900026d4924a10e02b871
|
| 3 |
+
size 44010583
|
checkpoints-v4.5/checkpoint-15360/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9418f8e34c62af4e9d503f15602f082908fcdeb9269e312cc231e43c0b80cc25
|
| 3 |
+
size 38935624
|
checkpoints-v4.5/checkpoint-15360/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5ea0878ca71cd118faee1f5e4b913522909d2cfafc0e42ae2892344300dc1c2b
|
| 3 |
+
size 77992907
|
checkpoints-v4.5/checkpoint-15360/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ed88f98b2f7c0b1c32b3c509e0d70558721a67865c66473a3f432aa221a1963c
|
| 3 |
+
size 14645
|
checkpoints-v4.5/checkpoint-15360/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e249e06856de6ebb0ada2baf7f2077a0f77d624c6d9f45bede75c169eeacf2d3
|
| 3 |
+
size 1383
|
checkpoints-v4.5/checkpoint-15360/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:73b7518420a541537ada4d348aaec8bae1a30b4c2513f08dc3817ea0a09a1265
|
| 3 |
+
size 1465
|
checkpoints-v4.5/checkpoint-15360/trainer_state.json
ADDED
|
@@ -0,0 +1,919 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.709436053761951,
|
| 6 |
+
"eval_steps": 1024,
|
| 7 |
+
"global_step": 15360,
|
| 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": 1.1487267017364502,
|
| 15 |
+
"learning_rate": 0.000498046875,
|
| 16 |
+
"loss": 11.798027992248535,
|
| 17 |
+
"step": 256
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.023647868458731697,
|
| 21 |
+
"grad_norm": 0.8386015295982361,
|
| 22 |
+
"learning_rate": 0.000998046875,
|
| 23 |
+
"loss": 1.7853779792785645,
|
| 24 |
+
"step": 512
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.03547180268809755,
|
| 28 |
+
"grad_norm": 0.7344363331794739,
|
| 29 |
+
"learning_rate": 0.000999640996023194,
|
| 30 |
+
"loss": 1.103014588356018,
|
| 31 |
+
"step": 768
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.047295736917463395,
|
| 35 |
+
"grad_norm": 1.1188315153121948,
|
| 36 |
+
"learning_rate": 0.0009985588674043958,
|
| 37 |
+
"loss": 0.9580796360969543,
|
| 38 |
+
"step": 1024
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.047295736917463395,
|
| 42 |
+
"eval_bleu": 0.9192375404443994,
|
| 43 |
+
"eval_ce_loss": 0.26167990953648745,
|
| 44 |
+
"eval_cos_loss": 0.26406568816127296,
|
| 45 |
+
"eval_loss": 0.9037111119864738,
|
| 46 |
+
"eval_mse_loss": 0.6016385473617135,
|
| 47 |
+
"eval_rec_loss": 0.013986090569199833,
|
| 48 |
+
"flow/cos_sim": 0.7359343250048215,
|
| 49 |
+
"flow/improvement_ratio": 0.9760946458605326,
|
| 50 |
+
"flow/mag_ratio_mean": 0.7269674200717717,
|
| 51 |
+
"flow/mag_ratio_std": 0.1390784539316343,
|
| 52 |
+
"step": 1024
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.047295736917463395,
|
| 56 |
+
"eval_bleu": 0.9192375404443994,
|
| 57 |
+
"eval_ce_loss": 0.26167990953648745,
|
| 58 |
+
"eval_cos_loss": 0.26406568816127296,
|
| 59 |
+
"eval_loss": 0.9037111119864738,
|
| 60 |
+
"eval_mse_loss": 0.6016385473617135,
|
| 61 |
+
"eval_rec_loss": 0.013986090569199833,
|
| 62 |
+
"eval_runtime": 144.0156,
|
| 63 |
+
"eval_samples_per_second": 194.375,
|
| 64 |
+
"eval_steps_per_second": 3.041,
|
| 65 |
+
"flow/cos_sim": 0.7359343250048215,
|
| 66 |
+
"flow/improvement_ratio": 0.9760946458605326,
|
| 67 |
+
"flow/mag_ratio_mean": 0.7269674200717717,
|
| 68 |
+
"flow/mag_ratio_std": 0.1390784539316343,
|
| 69 |
+
"step": 1024
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.05911967114682925,
|
| 73 |
+
"grad_norm": 1.0113043785095215,
|
| 74 |
+
"learning_rate": 0.0009967551747861387,
|
| 75 |
+
"loss": 0.8836896419525146,
|
| 76 |
+
"step": 1280
|
| 77 |
+
},
|
| 78 |
+
{
|
| 79 |
+
"epoch": 0.0709436053761951,
|
| 80 |
+
"grad_norm": 0.9680636525154114,
|
| 81 |
+
"learning_rate": 0.000994232528651847,
|
| 82 |
+
"loss": 0.8432819247245789,
|
| 83 |
+
"step": 1536
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.08276753960556095,
|
| 87 |
+
"grad_norm": 1.166627049446106,
|
| 88 |
+
"learning_rate": 0.0009909945800260092,
|
| 89 |
+
"loss": 0.7870283126831055,
|
| 90 |
+
"step": 1792
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 0.09459147383492679,
|
| 94 |
+
"grad_norm": 0.7747617363929749,
|
| 95 |
+
"learning_rate": 0.0009870460151900522,
|
| 96 |
+
"loss": 0.7735522389411926,
|
| 97 |
+
"step": 2048
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 0.09459147383492679,
|
| 101 |
+
"eval_bleu": 0.9264483676285591,
|
| 102 |
+
"eval_ce_loss": 0.22021640589690372,
|
| 103 |
+
"eval_cos_loss": 0.15934634184864557,
|
| 104 |
+
"eval_loss": 0.751849727815689,
|
| 105 |
+
"eval_mse_loss": 0.51190055103879,
|
| 106 |
+
"eval_rec_loss": 0.0037981332031229603,
|
| 107 |
+
"flow/cos_sim": 0.8406536742432477,
|
| 108 |
+
"flow/improvement_ratio": 0.9754726998337871,
|
| 109 |
+
"flow/mag_ratio_mean": 0.8395581131112085,
|
| 110 |
+
"flow/mag_ratio_std": 0.09706963221096013,
|
| 111 |
+
"step": 2048
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.09459147383492679,
|
| 115 |
+
"eval_bleu": 0.9264483676285591,
|
| 116 |
+
"eval_ce_loss": 0.22021640589690372,
|
| 117 |
+
"eval_cos_loss": 0.15934634184864557,
|
| 118 |
+
"eval_loss": 0.751849727815689,
|
| 119 |
+
"eval_mse_loss": 0.51190055103879,
|
| 120 |
+
"eval_rec_loss": 0.0037981332031229603,
|
| 121 |
+
"eval_runtime": 139.3758,
|
| 122 |
+
"eval_samples_per_second": 200.845,
|
| 123 |
+
"eval_steps_per_second": 3.143,
|
| 124 |
+
"flow/cos_sim": 0.8406536742432477,
|
| 125 |
+
"flow/improvement_ratio": 0.9754726998337871,
|
| 126 |
+
"flow/mag_ratio_mean": 0.8395581131112085,
|
| 127 |
+
"flow/mag_ratio_std": 0.09706963221096013,
|
| 128 |
+
"step": 2048
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.10641540806429264,
|
| 132 |
+
"grad_norm": 1.3507742881774902,
|
| 133 |
+
"learning_rate": 0.0009823925488998885,
|
| 134 |
+
"loss": 0.7531520128250122,
|
| 135 |
+
"step": 2304
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.1182393422936585,
|
| 139 |
+
"grad_norm": 1.090326189994812,
|
| 140 |
+
"learning_rate": 0.0009770409161149525,
|
| 141 |
+
"loss": 0.7384664416313171,
|
| 142 |
+
"step": 2560
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.13006327652302435,
|
| 146 |
+
"grad_norm": 1.6648627519607544,
|
| 147 |
+
"learning_rate": 0.0009709988622506973,
|
| 148 |
+
"loss": 0.7159472107887268,
|
| 149 |
+
"step": 2816
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.1418872107523902,
|
| 153 |
+
"grad_norm": 0.720405638217926,
|
| 154 |
+
"learning_rate": 0.000964275131968659,
|
| 155 |
+
"loss": 0.7134207487106323,
|
| 156 |
+
"step": 3072
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.1418872107523902,
|
| 160 |
+
"eval_bleu": 0.9362046037171927,
|
| 161 |
+
"eval_ce_loss": 0.18669388993094638,
|
| 162 |
+
"eval_cos_loss": 0.09599785676829892,
|
| 163 |
+
"eval_loss": 0.7009708360177741,
|
| 164 |
+
"eval_mse_loss": 0.5027340843797274,
|
| 165 |
+
"eval_rec_loss": 0.0019430727923829022,
|
| 166 |
+
"flow/cos_sim": 0.9040021625555814,
|
| 167 |
+
"flow/improvement_ratio": 0.9753212502830104,
|
| 168 |
+
"flow/mag_ratio_mean": 0.906250716208323,
|
| 169 |
+
"flow/mag_ratio_std": 0.07307879087519428,
|
| 170 |
+
"step": 3072
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.1418872107523902,
|
| 174 |
+
"eval_bleu": 0.9362046037171927,
|
| 175 |
+
"eval_ce_loss": 0.18669388993094638,
|
| 176 |
+
"eval_cos_loss": 0.09599785676829892,
|
| 177 |
+
"eval_loss": 0.7009708360177741,
|
| 178 |
+
"eval_mse_loss": 0.5027340843797274,
|
| 179 |
+
"eval_rec_loss": 0.0019430727923829022,
|
| 180 |
+
"eval_runtime": 139.2617,
|
| 181 |
+
"eval_samples_per_second": 201.01,
|
| 182 |
+
"eval_steps_per_second": 3.145,
|
| 183 |
+
"flow/cos_sim": 0.9040021625555814,
|
| 184 |
+
"flow/improvement_ratio": 0.9753212502830104,
|
| 185 |
+
"flow/mag_ratio_mean": 0.906250716208323,
|
| 186 |
+
"flow/mag_ratio_std": 0.07307879087519428,
|
| 187 |
+
"step": 3072
|
| 188 |
+
},
|
| 189 |
+
{
|
| 190 |
+
"epoch": 0.15371114498175603,
|
| 191 |
+
"grad_norm": 1.171225905418396,
|
| 192 |
+
"learning_rate": 0.0009568794565203123,
|
| 193 |
+
"loss": 0.6967981457710266,
|
| 194 |
+
"step": 3328
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.1655350792111219,
|
| 198 |
+
"grad_norm": 0.9184499979019165,
|
| 199 |
+
"learning_rate": 0.0009488225396630347,
|
| 200 |
+
"loss": 0.6859588623046875,
|
| 201 |
+
"step": 3584
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.17735901344048774,
|
| 205 |
+
"grad_norm": 1.0972322225570679,
|
| 206 |
+
"learning_rate": 0.0009401160421685646,
|
| 207 |
+
"loss": 0.68483966588974,
|
| 208 |
+
"step": 3840
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 0.18918294766985358,
|
| 212 |
+
"grad_norm": 1.2944236993789673,
|
| 213 |
+
"learning_rate": 0.0009307725649463714,
|
| 214 |
+
"loss": 0.6722217202186584,
|
| 215 |
+
"step": 4096
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 0.18918294766985358,
|
| 219 |
+
"eval_bleu": 0.9367529454304925,
|
| 220 |
+
"eval_ce_loss": 0.17851220300531687,
|
| 221 |
+
"eval_cos_loss": 0.0685597131088308,
|
| 222 |
+
"eval_loss": 0.6638682314522191,
|
| 223 |
+
"eval_mse_loss": 0.4771829223660029,
|
| 224 |
+
"eval_rec_loss": 0.0013171346048419428,
|
| 225 |
+
"flow/cos_sim": 0.9314403127045392,
|
| 226 |
+
"flow/improvement_ratio": 0.9747336862021929,
|
| 227 |
+
"flow/mag_ratio_mean": 0.9314602082722807,
|
| 228 |
+
"flow/mag_ratio_std": 0.05983651272068013,
|
| 229 |
+
"step": 4096
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 0.18918294766985358,
|
| 233 |
+
"eval_bleu": 0.9367529454304925,
|
| 234 |
+
"eval_ce_loss": 0.17851220300531687,
|
| 235 |
+
"eval_cos_loss": 0.0685597131088308,
|
| 236 |
+
"eval_loss": 0.6638682314522191,
|
| 237 |
+
"eval_mse_loss": 0.4771829223660029,
|
| 238 |
+
"eval_rec_loss": 0.0013171346048419428,
|
| 239 |
+
"eval_runtime": 143.247,
|
| 240 |
+
"eval_samples_per_second": 195.418,
|
| 241 |
+
"eval_steps_per_second": 3.058,
|
| 242 |
+
"flow/cos_sim": 0.9314403127045392,
|
| 243 |
+
"flow/improvement_ratio": 0.9747336862021929,
|
| 244 |
+
"flow/mag_ratio_mean": 0.9314602082722807,
|
| 245 |
+
"flow/mag_ratio_std": 0.05983651272068013,
|
| 246 |
+
"step": 4096
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 0.20100688189921945,
|
| 250 |
+
"grad_norm": 1.0950664281845093,
|
| 251 |
+
"learning_rate": 0.0009208056308063659,
|
| 252 |
+
"loss": 0.6635431051254272,
|
| 253 |
+
"step": 4352
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 0.2128308161285853,
|
| 257 |
+
"grad_norm": 1.0510311126708984,
|
| 258 |
+
"learning_rate": 0.0009102296648873445,
|
| 259 |
+
"loss": 0.652130126953125,
|
| 260 |
+
"step": 4608
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.22465475035795113,
|
| 264 |
+
"grad_norm": 0.7107524275779724,
|
| 265 |
+
"learning_rate": 0.0008990599737794927,
|
| 266 |
+
"loss": 0.6548014283180237,
|
| 267 |
+
"step": 4864
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.236478684587317,
|
| 271 |
+
"grad_norm": 1.119279146194458,
|
| 272 |
+
"learning_rate": 0.0008873127233711644,
|
| 273 |
+
"loss": 0.644295871257782,
|
| 274 |
+
"step": 5120
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 0.236478684587317,
|
| 278 |
+
"eval_bleu": 0.9375087809294717,
|
| 279 |
+
"eval_ce_loss": 0.17866700632629498,
|
| 280 |
+
"eval_cos_loss": 0.05568918508379699,
|
| 281 |
+
"eval_loss": 0.6384022356304404,
|
| 282 |
+
"eval_mse_loss": 0.4531491862856634,
|
| 283 |
+
"eval_rec_loss": 0.0010171248973892112,
|
| 284 |
+
"flow/cos_sim": 0.9443108406785417,
|
| 285 |
+
"flow/improvement_ratio": 0.9754358607612245,
|
| 286 |
+
"flow/mag_ratio_mean": 0.9432625220790846,
|
| 287 |
+
"flow/mag_ratio_std": 0.0532344374550532,
|
| 288 |
+
"step": 5120
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 0.236478684587317,
|
| 292 |
+
"eval_bleu": 0.9375087809294717,
|
| 293 |
+
"eval_ce_loss": 0.17866700632629498,
|
| 294 |
+
"eval_cos_loss": 0.05568918508379699,
|
| 295 |
+
"eval_loss": 0.6384022356304404,
|
| 296 |
+
"eval_mse_loss": 0.4531491862856634,
|
| 297 |
+
"eval_rec_loss": 0.0010171248973892112,
|
| 298 |
+
"eval_runtime": 140.7376,
|
| 299 |
+
"eval_samples_per_second": 198.902,
|
| 300 |
+
"eval_steps_per_second": 3.112,
|
| 301 |
+
"flow/cos_sim": 0.9443108406785417,
|
| 302 |
+
"flow/improvement_ratio": 0.9754358607612245,
|
| 303 |
+
"flow/mag_ratio_mean": 0.9432625220790846,
|
| 304 |
+
"flow/mag_ratio_std": 0.0532344374550532,
|
| 305 |
+
"step": 5120
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 0.24830261881668284,
|
| 309 |
+
"grad_norm": 1.223329782485962,
|
| 310 |
+
"learning_rate": 0.0008750049154520011,
|
| 311 |
+
"loss": 0.6385497450828552,
|
| 312 |
+
"step": 5376
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 0.2601265530460487,
|
| 316 |
+
"grad_norm": 0.7951129078865051,
|
| 317 |
+
"learning_rate": 0.0008621543631062487,
|
| 318 |
+
"loss": 0.6350575089454651,
|
| 319 |
+
"step": 5632
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 0.27195048727541454,
|
| 323 |
+
"grad_norm": 0.8830247521400452,
|
| 324 |
+
"learning_rate": 0.0008487796649318904,
|
| 325 |
+
"loss": 0.6269800066947937,
|
| 326 |
+
"step": 5888
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.2837744215047804,
|
| 330 |
+
"grad_norm": 1.0399079322814941,
|
| 331 |
+
"learning_rate": 0.0008349001781229053,
|
| 332 |
+
"loss": 0.6236906051635742,
|
| 333 |
+
"step": 6144
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.2837744215047804,
|
| 337 |
+
"eval_bleu": 0.9406288888554537,
|
| 338 |
+
"eval_ce_loss": 0.16271350685439018,
|
| 339 |
+
"eval_cos_loss": 0.04946032597696128,
|
| 340 |
+
"eval_loss": 0.6158577807962078,
|
| 341 |
+
"eval_mse_loss": 0.4472781540868489,
|
| 342 |
+
"eval_rec_loss": 0.0009200886164281037,
|
| 343 |
+
"flow/cos_sim": 0.9505396935765602,
|
| 344 |
+
"flow/improvement_ratio": 0.9752184091365501,
|
| 345 |
+
"flow/mag_ratio_mean": 0.9574754819205907,
|
| 346 |
+
"flow/mag_ratio_std": 0.04891788842131014,
|
| 347 |
+
"step": 6144
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 0.2837744215047804,
|
| 351 |
+
"eval_bleu": 0.9406288888554537,
|
| 352 |
+
"eval_ce_loss": 0.16271350685439018,
|
| 353 |
+
"eval_cos_loss": 0.04946032597696128,
|
| 354 |
+
"eval_loss": 0.6158577807962078,
|
| 355 |
+
"eval_mse_loss": 0.4472781540868489,
|
| 356 |
+
"eval_rec_loss": 0.0009200886164281037,
|
| 357 |
+
"eval_runtime": 138.7881,
|
| 358 |
+
"eval_samples_per_second": 201.696,
|
| 359 |
+
"eval_steps_per_second": 3.156,
|
| 360 |
+
"flow/cos_sim": 0.9505396935765602,
|
| 361 |
+
"flow/improvement_ratio": 0.9752184091365501,
|
| 362 |
+
"flow/mag_ratio_mean": 0.9574754819205907,
|
| 363 |
+
"flow/mag_ratio_std": 0.04891788842131014,
|
| 364 |
+
"step": 6144
|
| 365 |
+
},
|
| 366 |
+
{
|
| 367 |
+
"epoch": 0.2955983557341462,
|
| 368 |
+
"grad_norm": 1.2804330587387085,
|
| 369 |
+
"learning_rate": 0.0008205359904536107,
|
| 370 |
+
"loss": 0.6104704737663269,
|
| 371 |
+
"step": 6400
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 0.30742228996351206,
|
| 375 |
+
"grad_norm": 1.038807988166809,
|
| 376 |
+
"learning_rate": 0.0008057078912056363,
|
| 377 |
+
"loss": 0.6035579442977905,
|
| 378 |
+
"step": 6656
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 0.3192462241928779,
|
| 382 |
+
"grad_norm": 1.1162539720535278,
|
| 383 |
+
"learning_rate": 0.0007904373410796086,
|
| 384 |
+
"loss": 0.6099694967269897,
|
| 385 |
+
"step": 6912
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 0.3310701584222438,
|
| 389 |
+
"grad_norm": 0.8053554892539978,
|
| 390 |
+
"learning_rate": 0.0007747464411350876,
|
| 391 |
+
"loss": 0.6010444760322571,
|
| 392 |
+
"step": 7168
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.3310701584222438,
|
| 396 |
+
"eval_bleu": 0.9403609007223014,
|
| 397 |
+
"eval_ce_loss": 0.16489822024130793,
|
| 398 |
+
"eval_cos_loss": 0.042615741474307293,
|
| 399 |
+
"eval_loss": 0.5965314044799979,
|
| 400 |
+
"eval_mse_loss": 0.4266453656839998,
|
| 401 |
+
"eval_rec_loss": 0.0007262465627901213,
|
| 402 |
+
"flow/cos_sim": 0.9573842790573155,
|
| 403 |
+
"flow/improvement_ratio": 0.9754182146564466,
|
| 404 |
+
"flow/mag_ratio_mean": 0.9583017167435389,
|
| 405 |
+
"flow/mag_ratio_std": 0.044361623012584096,
|
| 406 |
+
"step": 7168
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 0.3310701584222438,
|
| 410 |
+
"eval_bleu": 0.9403609007223014,
|
| 411 |
+
"eval_ce_loss": 0.16489822024130793,
|
| 412 |
+
"eval_cos_loss": 0.042615741474307293,
|
| 413 |
+
"eval_loss": 0.5965314044799979,
|
| 414 |
+
"eval_mse_loss": 0.4266453656839998,
|
| 415 |
+
"eval_rec_loss": 0.0007262465627901213,
|
| 416 |
+
"eval_runtime": 141.0373,
|
| 417 |
+
"eval_samples_per_second": 198.479,
|
| 418 |
+
"eval_steps_per_second": 3.106,
|
| 419 |
+
"flow/cos_sim": 0.9573842790573155,
|
| 420 |
+
"flow/improvement_ratio": 0.9754182146564466,
|
| 421 |
+
"flow/mag_ratio_mean": 0.9583017167435389,
|
| 422 |
+
"flow/mag_ratio_std": 0.044361623012584096,
|
| 423 |
+
"step": 7168
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.34289409265160964,
|
| 427 |
+
"grad_norm": 0.926774799823761,
|
| 428 |
+
"learning_rate": 0.000758657900803716,
|
| 429 |
+
"loss": 0.6045265793800354,
|
| 430 |
+
"step": 7424
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.3547180268809755,
|
| 434 |
+
"grad_norm": 0.6091651320457458,
|
| 435 |
+
"learning_rate": 0.000742195005021869,
|
| 436 |
+
"loss": 0.6046400666236877,
|
| 437 |
+
"step": 7680
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.3665419611103413,
|
| 441 |
+
"grad_norm": 0.9995866417884827,
|
| 442 |
+
"learning_rate": 0.0007253815805303786,
|
| 443 |
+
"loss": 0.5923656225204468,
|
| 444 |
+
"step": 7936
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.37836589533970716,
|
| 448 |
+
"grad_norm": 0.8947123885154724,
|
| 449 |
+
"learning_rate": 0.0007082419613901028,
|
| 450 |
+
"loss": 0.5886037349700928,
|
| 451 |
+
"step": 8192
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.37836589533970716,
|
| 455 |
+
"eval_bleu": 0.9428308469149022,
|
| 456 |
+
"eval_ce_loss": 0.15756705103943883,
|
| 457 |
+
"eval_cos_loss": 0.03951380795641849,
|
| 458 |
+
"eval_loss": 0.5839528071281572,
|
| 459 |
+
"eval_mse_loss": 0.42179430878325685,
|
| 460 |
+
"eval_rec_loss": 0.0006400666907252946,
|
| 461 |
+
"flow/cos_sim": 0.9604862126857723,
|
| 462 |
+
"flow/improvement_ratio": 0.9746591260988419,
|
| 463 |
+
"flow/mag_ratio_mean": 0.9592139723638421,
|
| 464 |
+
"flow/mag_ratio_std": 0.04008094550505893,
|
| 465 |
+
"step": 8192
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.37836589533970716,
|
| 469 |
+
"eval_bleu": 0.9428308469149022,
|
| 470 |
+
"eval_ce_loss": 0.15756705103943883,
|
| 471 |
+
"eval_cos_loss": 0.03951380795641849,
|
| 472 |
+
"eval_loss": 0.5839528071281572,
|
| 473 |
+
"eval_mse_loss": 0.42179430878325685,
|
| 474 |
+
"eval_rec_loss": 0.0006400666907252946,
|
| 475 |
+
"eval_runtime": 142.8618,
|
| 476 |
+
"eval_samples_per_second": 195.945,
|
| 477 |
+
"eval_steps_per_second": 3.066,
|
| 478 |
+
"flow/cos_sim": 0.9604862126857723,
|
| 479 |
+
"flow/improvement_ratio": 0.9746591260988419,
|
| 480 |
+
"flow/mag_ratio_mean": 0.9592139723638421,
|
| 481 |
+
"flow/mag_ratio_std": 0.04008094550505893,
|
| 482 |
+
"step": 8192
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.390189829569073,
|
| 486 |
+
"grad_norm": 0.9762691259384155,
|
| 487 |
+
"learning_rate": 0.0006908009537632514,
|
| 488 |
+
"loss": 0.591374397277832,
|
| 489 |
+
"step": 8448
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.4020137637984389,
|
| 493 |
+
"grad_norm": 0.9119466543197632,
|
| 494 |
+
"learning_rate": 0.0006730838000114403,
|
| 495 |
+
"loss": 0.5883693099021912,
|
| 496 |
+
"step": 8704
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 0.41383769802780473,
|
| 500 |
+
"grad_norm": 0.7088457942008972,
|
| 501 |
+
"learning_rate": 0.0006551161421624341,
|
| 502 |
+
"loss": 0.591505229473114,
|
| 503 |
+
"step": 8960
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 0.4256616322571706,
|
| 507 |
+
"grad_norm": 0.962149441242218,
|
| 508 |
+
"learning_rate": 0.0006369239847984517,
|
| 509 |
+
"loss": 0.5791484117507935,
|
| 510 |
+
"step": 9216
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 0.4256616322571706,
|
| 514 |
+
"eval_bleu": 0.9397317595032229,
|
| 515 |
+
"eval_ce_loss": 0.1654567693755643,
|
| 516 |
+
"eval_cos_loss": 0.036672262636493876,
|
| 517 |
+
"eval_loss": 0.5844078893394775,
|
| 518 |
+
"eval_mse_loss": 0.41475171422305174,
|
| 519 |
+
"eval_rec_loss": 0.0005321793435803258,
|
| 520 |
+
"flow/cos_sim": 0.9633277582523485,
|
| 521 |
+
"flow/improvement_ratio": 0.9749675198504913,
|
| 522 |
+
"flow/mag_ratio_mean": 0.9673667468436776,
|
| 523 |
+
"flow/mag_ratio_std": 0.03610456871829893,
|
| 524 |
+
"step": 9216
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.4256616322571706,
|
| 528 |
+
"eval_bleu": 0.9397317595032229,
|
| 529 |
+
"eval_ce_loss": 0.1654567693755643,
|
| 530 |
+
"eval_cos_loss": 0.036672262636493876,
|
| 531 |
+
"eval_loss": 0.5844078893394775,
|
| 532 |
+
"eval_mse_loss": 0.41475171422305174,
|
| 533 |
+
"eval_rec_loss": 0.0005321793435803258,
|
| 534 |
+
"eval_runtime": 140.1978,
|
| 535 |
+
"eval_samples_per_second": 199.668,
|
| 536 |
+
"eval_steps_per_second": 3.124,
|
| 537 |
+
"flow/cos_sim": 0.9633277582523485,
|
| 538 |
+
"flow/improvement_ratio": 0.9749675198504913,
|
| 539 |
+
"flow/mag_ratio_mean": 0.9673667468436776,
|
| 540 |
+
"flow/mag_ratio_std": 0.03610456871829893,
|
| 541 |
+
"step": 9216
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.4374855664865364,
|
| 545 |
+
"grad_norm": 0.853702187538147,
|
| 546 |
+
"learning_rate": 0.0006185336574197479,
|
| 547 |
+
"loss": 0.5742002725601196,
|
| 548 |
+
"step": 9472
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.44930950071590225,
|
| 552 |
+
"grad_norm": 0.9982457160949707,
|
| 553 |
+
"learning_rate": 0.0005999717763379407,
|
| 554 |
+
"loss": 0.5812740921974182,
|
| 555 |
+
"step": 9728
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.4611334349452681,
|
| 559 |
+
"grad_norm": 1.16475510597229,
|
| 560 |
+
"learning_rate": 0.0005812652061542363,
|
| 561 |
+
"loss": 0.5733819603919983,
|
| 562 |
+
"step": 9984
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.472957369174634,
|
| 566 |
+
"grad_norm": 1.0706837177276611,
|
| 567 |
+
"learning_rate": 0.0005624410208783071,
|
| 568 |
+
"loss": 0.5738887190818787,
|
| 569 |
+
"step": 10240
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.472957369174634,
|
| 573 |
+
"eval_bleu": 0.9459802582441368,
|
| 574 |
+
"eval_ce_loss": 0.14516126815459296,
|
| 575 |
+
"eval_cos_loss": 0.033954130285780995,
|
| 576 |
+
"eval_loss": 0.5643789048336413,
|
| 577 |
+
"eval_mse_loss": 0.4153396790702593,
|
| 578 |
+
"eval_rec_loss": 0.0004825431066599768,
|
| 579 |
+
"flow/cos_sim": 0.9660458944431723,
|
| 580 |
+
"flow/improvement_ratio": 0.9768404536051293,
|
| 581 |
+
"flow/mag_ratio_mean": 0.964692687198996,
|
| 582 |
+
"flow/mag_ratio_std": 0.033606582049059266,
|
| 583 |
+
"step": 10240
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.472957369174634,
|
| 587 |
+
"eval_bleu": 0.9459802582441368,
|
| 588 |
+
"eval_ce_loss": 0.14516126815459296,
|
| 589 |
+
"eval_cos_loss": 0.033954130285780995,
|
| 590 |
+
"eval_loss": 0.5643789048336413,
|
| 591 |
+
"eval_mse_loss": 0.4153396790702593,
|
| 592 |
+
"eval_rec_loss": 0.0004825431066599768,
|
| 593 |
+
"eval_runtime": 138.7003,
|
| 594 |
+
"eval_samples_per_second": 201.824,
|
| 595 |
+
"eval_steps_per_second": 3.158,
|
| 596 |
+
"flow/cos_sim": 0.9660458944431723,
|
| 597 |
+
"flow/improvement_ratio": 0.9768404536051293,
|
| 598 |
+
"flow/mag_ratio_mean": 0.964692687198996,
|
| 599 |
+
"flow/mag_ratio_std": 0.033606582049059266,
|
| 600 |
+
"step": 10240
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 0.48478130340399983,
|
| 604 |
+
"grad_norm": 1.2897191047668457,
|
| 605 |
+
"learning_rate": 0.0005435264647440881,
|
| 606 |
+
"loss": 0.5754253268241882,
|
| 607 |
+
"step": 10496
|
| 608 |
+
},
|
| 609 |
+
{
|
| 610 |
+
"epoch": 0.49660523763336567,
|
| 611 |
+
"grad_norm": 0.8889420628547668,
|
| 612 |
+
"learning_rate": 0.000524548912779213,
|
| 613 |
+
"loss": 0.5645499229431152,
|
| 614 |
+
"step": 10752
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 0.5084291718627315,
|
| 618 |
+
"grad_norm": 1.0012623071670532,
|
| 619 |
+
"learning_rate": 0.0005055358311851499,
|
| 620 |
+
"loss": 0.570586085319519,
|
| 621 |
+
"step": 11008
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 0.5202531060920974,
|
| 625 |
+
"grad_norm": 0.7268548011779785,
|
| 626 |
+
"learning_rate": 0.0004865147375853812,
|
| 627 |
+
"loss": 0.5649828910827637,
|
| 628 |
+
"step": 11264
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"epoch": 0.5202531060920974,
|
| 632 |
+
"eval_bleu": 0.9438391043268173,
|
| 633 |
+
"eval_ce_loss": 0.15168385392003883,
|
| 634 |
+
"eval_cos_loss": 0.031372046842320596,
|
| 635 |
+
"eval_loss": 0.5559442507349737,
|
| 636 |
+
"eval_mse_loss": 0.40068320514948946,
|
| 637 |
+
"eval_rec_loss": 0.0004399877304908156,
|
| 638 |
+
"flow/cos_sim": 0.9686279725538541,
|
| 639 |
+
"flow/improvement_ratio": 0.9761811211773249,
|
| 640 |
+
"flow/mag_ratio_mean": 0.9699762897676529,
|
| 641 |
+
"flow/mag_ratio_std": 0.031497096553546926,
|
| 642 |
+
"step": 11264
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 0.5202531060920974,
|
| 646 |
+
"eval_bleu": 0.9438391043268173,
|
| 647 |
+
"eval_ce_loss": 0.15168385392003883,
|
| 648 |
+
"eval_cos_loss": 0.031372046842320596,
|
| 649 |
+
"eval_loss": 0.5559442507349737,
|
| 650 |
+
"eval_mse_loss": 0.40068320514948946,
|
| 651 |
+
"eval_rec_loss": 0.0004399877304908156,
|
| 652 |
+
"eval_runtime": 142.7289,
|
| 653 |
+
"eval_samples_per_second": 196.127,
|
| 654 |
+
"eval_steps_per_second": 3.069,
|
| 655 |
+
"flow/cos_sim": 0.9686279725538541,
|
| 656 |
+
"flow/improvement_ratio": 0.9761811211773249,
|
| 657 |
+
"flow/mag_ratio_mean": 0.9699762897676529,
|
| 658 |
+
"flow/mag_ratio_std": 0.031497096553546926,
|
| 659 |
+
"step": 11264
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 0.5320770403214632,
|
| 663 |
+
"grad_norm": 1.215234637260437,
|
| 664 |
+
"learning_rate": 0.0004675131611991607,
|
| 665 |
+
"loss": 0.5652971863746643,
|
| 666 |
+
"step": 11520
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 0.5439009745508291,
|
| 670 |
+
"grad_norm": 0.622643768787384,
|
| 671 |
+
"learning_rate": 0.0004485586029984899,
|
| 672 |
+
"loss": 0.5565246343612671,
|
| 673 |
+
"step": 11776
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 0.5557249087801949,
|
| 677 |
+
"grad_norm": 1.0961378812789917,
|
| 678 |
+
"learning_rate": 0.00042967849590597266,
|
| 679 |
+
"loss": 0.5489715933799744,
|
| 680 |
+
"step": 12032
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.5675488430095608,
|
| 684 |
+
"grad_norm": 0.9782881140708923,
|
| 685 |
+
"learning_rate": 0.0004109001650911621,
|
| 686 |
+
"loss": 0.5525087118148804,
|
| 687 |
+
"step": 12288
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 0.5675488430095608,
|
| 691 |
+
"eval_bleu": 0.9412862694661508,
|
| 692 |
+
"eval_ce_loss": 0.1635204538911343,
|
| 693 |
+
"eval_cos_loss": 0.030069323277874895,
|
| 694 |
+
"eval_loss": 0.5582960859689539,
|
| 695 |
+
"eval_mse_loss": 0.39135666776737665,
|
| 696 |
+
"eval_rec_loss": 0.00041203244906421176,
|
| 697 |
+
"flow/cos_sim": 0.969930695344324,
|
| 698 |
+
"flow/improvement_ratio": 0.9753037164472553,
|
| 699 |
+
"flow/mag_ratio_mean": 0.9802452540833112,
|
| 700 |
+
"flow/mag_ratio_std": 0.030909983262623827,
|
| 701 |
+
"step": 12288
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 0.5675488430095608,
|
| 705 |
+
"eval_bleu": 0.9412862694661508,
|
| 706 |
+
"eval_ce_loss": 0.1635204538911343,
|
| 707 |
+
"eval_cos_loss": 0.030069323277874895,
|
| 708 |
+
"eval_loss": 0.5582960859689539,
|
| 709 |
+
"eval_mse_loss": 0.39135666776737665,
|
| 710 |
+
"eval_rec_loss": 0.00041203244906421176,
|
| 711 |
+
"eval_runtime": 141.5788,
|
| 712 |
+
"eval_samples_per_second": 197.72,
|
| 713 |
+
"eval_steps_per_second": 3.094,
|
| 714 |
+
"flow/cos_sim": 0.969930695344324,
|
| 715 |
+
"flow/improvement_ratio": 0.9753037164472553,
|
| 716 |
+
"flow/mag_ratio_mean": 0.9802452540833112,
|
| 717 |
+
"flow/mag_ratio_std": 0.030909983262623827,
|
| 718 |
+
"step": 12288
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 0.5793727772389267,
|
| 722 |
+
"grad_norm": 1.1210603713989258,
|
| 723 |
+
"learning_rate": 0.0003922507884228551,
|
| 724 |
+
"loss": 0.5510575175285339,
|
| 725 |
+
"step": 12544
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 0.5911967114682924,
|
| 729 |
+
"grad_norm": 0.7037401795387268,
|
| 730 |
+
"learning_rate": 0.00037375735713457723,
|
| 731 |
+
"loss": 0.5481619238853455,
|
| 732 |
+
"step": 12800
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 0.6030206456976583,
|
| 736 |
+
"grad_norm": 0.7474592924118042,
|
| 737 |
+
"learning_rate": 0.00035544663676018276,
|
| 738 |
+
"loss": 0.5506067872047424,
|
| 739 |
+
"step": 13056
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 0.6148445799270241,
|
| 743 |
+
"grad_norm": 0.7188865542411804,
|
| 744 |
+
"learning_rate": 0.00033734512839611255,
|
| 745 |
+
"loss": 0.5451632738113403,
|
| 746 |
+
"step": 13312
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.6148445799270241,
|
| 750 |
+
"eval_bleu": 0.9428834105382213,
|
| 751 |
+
"eval_ce_loss": 0.1552433374967239,
|
| 752 |
+
"eval_cos_loss": 0.02893113123771807,
|
| 753 |
+
"eval_loss": 0.5456167366951024,
|
| 754 |
+
"eval_mse_loss": 0.3870963177316265,
|
| 755 |
+
"eval_rec_loss": 0.0003839662346667844,
|
| 756 |
+
"flow/cos_sim": 0.9710688916243375,
|
| 757 |
+
"flow/improvement_ratio": 0.9738836331998921,
|
| 758 |
+
"flow/mag_ratio_mean": 0.9706440245451993,
|
| 759 |
+
"flow/mag_ratio_std": 0.029164127312328446,
|
| 760 |
+
"step": 13312
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 0.6148445799270241,
|
| 764 |
+
"eval_bleu": 0.9428834105382213,
|
| 765 |
+
"eval_ce_loss": 0.1552433374967239,
|
| 766 |
+
"eval_cos_loss": 0.02893113123771807,
|
| 767 |
+
"eval_loss": 0.5456167366951024,
|
| 768 |
+
"eval_mse_loss": 0.3870963177316265,
|
| 769 |
+
"eval_rec_loss": 0.0003839662346667844,
|
| 770 |
+
"eval_runtime": 141.1326,
|
| 771 |
+
"eval_samples_per_second": 198.345,
|
| 772 |
+
"eval_steps_per_second": 3.103,
|
| 773 |
+
"flow/cos_sim": 0.9710688916243375,
|
| 774 |
+
"flow/improvement_ratio": 0.9738836331998921,
|
| 775 |
+
"flow/mag_ratio_mean": 0.9706440245451993,
|
| 776 |
+
"flow/mag_ratio_std": 0.029164127312328446,
|
| 777 |
+
"step": 13312
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 0.62666851415639,
|
| 781 |
+
"grad_norm": 0.7716706991195679,
|
| 782 |
+
"learning_rate": 0.0003194790303463687,
|
| 783 |
+
"loss": 0.537281334400177,
|
| 784 |
+
"step": 13568
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 0.6384924483857558,
|
| 788 |
+
"grad_norm": 1.332189917564392,
|
| 789 |
+
"learning_rate": 0.00030187420020572406,
|
| 790 |
+
"loss": 0.5493588447570801,
|
| 791 |
+
"step": 13824
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 0.6503163826151217,
|
| 795 |
+
"grad_norm": 0.7042582035064697,
|
| 796 |
+
"learning_rate": 0.00028455611743603626,
|
| 797 |
+
"loss": 0.5357389450073242,
|
| 798 |
+
"step": 14080
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 0.6621403168444876,
|
| 802 |
+
"grad_norm": 0.9820289611816406,
|
| 803 |
+
"learning_rate": 0.0002675498464898373,
|
| 804 |
+
"loss": 0.5378322601318359,
|
| 805 |
+
"step": 14336
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 0.6621403168444876,
|
| 809 |
+
"eval_bleu": 0.9466625254289972,
|
| 810 |
+
"eval_ce_loss": 0.14409655002562421,
|
| 811 |
+
"eval_cos_loss": 0.028247294284097137,
|
| 812 |
+
"eval_loss": 0.5304203147757544,
|
| 813 |
+
"eval_mse_loss": 0.38314900074375274,
|
| 814 |
+
"eval_rec_loss": 0.0003500353127845417,
|
| 815 |
+
"flow/cos_sim": 0.9717527268684074,
|
| 816 |
+
"flow/improvement_ratio": 0.9754393091212669,
|
| 817 |
+
"flow/mag_ratio_mean": 0.973318548915593,
|
| 818 |
+
"flow/mag_ratio_std": 0.028512526988200674,
|
| 819 |
+
"step": 14336
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.6621403168444876,
|
| 823 |
+
"eval_bleu": 0.9466625254289972,
|
| 824 |
+
"eval_ce_loss": 0.14409655002562421,
|
| 825 |
+
"eval_cos_loss": 0.028247294284097137,
|
| 826 |
+
"eval_loss": 0.5304203147757544,
|
| 827 |
+
"eval_mse_loss": 0.38314900074375274,
|
| 828 |
+
"eval_rec_loss": 0.0003500353127845417,
|
| 829 |
+
"eval_runtime": 142.684,
|
| 830 |
+
"eval_samples_per_second": 196.189,
|
| 831 |
+
"eval_steps_per_second": 3.07,
|
| 832 |
+
"flow/cos_sim": 0.9717527268684074,
|
| 833 |
+
"flow/improvement_ratio": 0.9754393091212669,
|
| 834 |
+
"flow/mag_ratio_mean": 0.973318548915593,
|
| 835 |
+
"flow/mag_ratio_std": 0.028512526988200674,
|
| 836 |
+
"step": 14336
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.6739642510738534,
|
| 840 |
+
"grad_norm": 0.9396564364433289,
|
| 841 |
+
"learning_rate": 0.0002508800005345623,
|
| 842 |
+
"loss": 0.5384619832038879,
|
| 843 |
+
"step": 14592
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.6857881853032193,
|
| 847 |
+
"grad_norm": 1.3034203052520752,
|
| 848 |
+
"learning_rate": 0.00023457070582992562,
|
| 849 |
+
"loss": 0.5381016135215759,
|
| 850 |
+
"step": 14848
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 0.6976121195325851,
|
| 854 |
+
"grad_norm": 0.7913278341293335,
|
| 855 |
+
"learning_rate": 0.00021864556680999692,
|
| 856 |
+
"loss": 0.5290021896362305,
|
| 857 |
+
"step": 15104
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 0.709436053761951,
|
| 861 |
+
"grad_norm": 0.9535462856292725,
|
| 862 |
+
"learning_rate": 0.0002031276319205152,
|
| 863 |
+
"loss": 0.534302830696106,
|
| 864 |
+
"step": 15360
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 0.709436053761951,
|
| 868 |
+
"eval_bleu": 0.9467887438059129,
|
| 869 |
+
"eval_ce_loss": 0.14362057716241233,
|
| 870 |
+
"eval_cos_loss": 0.02781058812607505,
|
| 871 |
+
"eval_loss": 0.5298489093372266,
|
| 872 |
+
"eval_mse_loss": 0.38311818289702343,
|
| 873 |
+
"eval_rec_loss": 0.00032909089600148805,
|
| 874 |
+
"flow/cos_sim": 0.9721894303685454,
|
| 875 |
+
"flow/improvement_ratio": 0.974593520844908,
|
| 876 |
+
"flow/mag_ratio_mean": 0.9698141316572825,
|
| 877 |
+
"flow/mag_ratio_std": 0.02832854554465372,
|
| 878 |
+
"step": 15360
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.709436053761951,
|
| 882 |
+
"eval_bleu": 0.9467887438059129,
|
| 883 |
+
"eval_ce_loss": 0.14362057716241233,
|
| 884 |
+
"eval_cos_loss": 0.02781058812607505,
|
| 885 |
+
"eval_loss": 0.5298489093372266,
|
| 886 |
+
"eval_mse_loss": 0.38311818289702343,
|
| 887 |
+
"eval_rec_loss": 0.00032909089600148805,
|
| 888 |
+
"eval_runtime": 141.9779,
|
| 889 |
+
"eval_samples_per_second": 197.164,
|
| 890 |
+
"eval_steps_per_second": 3.085,
|
| 891 |
+
"flow/cos_sim": 0.9721894303685454,
|
| 892 |
+
"flow/improvement_ratio": 0.974593520844908,
|
| 893 |
+
"flow/mag_ratio_mean": 0.9698141316572825,
|
| 894 |
+
"flow/mag_ratio_std": 0.02832854554465372,
|
| 895 |
+
"step": 15360
|
| 896 |
+
}
|
| 897 |
+
],
|
| 898 |
+
"logging_steps": 256,
|
| 899 |
+
"max_steps": 21651,
|
| 900 |
+
"num_input_tokens_seen": 0,
|
| 901 |
+
"num_train_epochs": 1,
|
| 902 |
+
"save_steps": 1024,
|
| 903 |
+
"stateful_callbacks": {
|
| 904 |
+
"TrainerControl": {
|
| 905 |
+
"args": {
|
| 906 |
+
"should_epoch_stop": false,
|
| 907 |
+
"should_evaluate": false,
|
| 908 |
+
"should_log": false,
|
| 909 |
+
"should_save": true,
|
| 910 |
+
"should_training_stop": false
|
| 911 |
+
},
|
| 912 |
+
"attributes": {}
|
| 913 |
+
}
|
| 914 |
+
},
|
| 915 |
+
"total_flos": 0.0,
|
| 916 |
+
"train_batch_size": 64,
|
| 917 |
+
"trial_name": null,
|
| 918 |
+
"trial_params": null
|
| 919 |
+
}
|
checkpoints-v4.5/checkpoint-15360/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ee9cb7ebeac20164f12c8a5bf77bbe01722a7871151494aba288a393bf7de081
|
| 3 |
+
size 5137
|