Training in progress, step 18500, checkpoint
Browse files
last-checkpoint/adapter_config.json
CHANGED
|
@@ -29,13 +29,13 @@
|
|
| 29 |
"rank_pattern": {},
|
| 30 |
"revision": null,
|
| 31 |
"target_modules": [
|
| 32 |
-
"k_proj",
|
| 33 |
-
"up_proj",
|
| 34 |
-
"o_proj",
|
| 35 |
"v_proj",
|
| 36 |
"down_proj",
|
| 37 |
-
"
|
| 38 |
-
"
|
|
|
|
|
|
|
|
|
|
| 39 |
],
|
| 40 |
"target_parameters": null,
|
| 41 |
"task_type": "CAUSAL_LM",
|
|
|
|
| 29 |
"rank_pattern": {},
|
| 30 |
"revision": null,
|
| 31 |
"target_modules": [
|
|
|
|
|
|
|
|
|
|
| 32 |
"v_proj",
|
| 33 |
"down_proj",
|
| 34 |
+
"o_proj",
|
| 35 |
+
"k_proj",
|
| 36 |
+
"q_proj",
|
| 37 |
+
"up_proj",
|
| 38 |
+
"gate_proj"
|
| 39 |
],
|
| 40 |
"target_parameters": null,
|
| 41 |
"task_type": "CAUSAL_LM",
|
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 295488936
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fe6bc82d0082aea807a8cf60117db13e55bce8c0365e985f5e323bd7d91e8e79
|
| 3 |
size 295488936
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 591209083
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6974dd099a342d4edb4399a77f947c011c69080b579f4ab294b43c3bdd45a95e
|
| 3 |
size 591209083
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14645
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cfdead6fb89febd0ec3fe4773ba2ee772b5ca6734d6e27d5e4932e94bc72716b
|
| 3 |
size 14645
|
last-checkpoint/scaler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1383
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0d1666da2e98dd1d6c433862a19b349ce96485574b7ec6f9231ed31994dccfa5
|
| 3 |
size 1383
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1465
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cd64daf4a3291c591a85097ad249508171725eaecaced914df4d2daf4e68a8e7
|
| 3 |
size 1465
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 1.
|
| 6 |
"eval_steps": 500,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -12896,6 +12896,364 @@
|
|
| 12896 |
"eval_samples_per_second": 2.929,
|
| 12897 |
"eval_steps_per_second": 2.929,
|
| 12898 |
"step": 18000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12899 |
}
|
| 12900 |
],
|
| 12901 |
"logging_steps": 10,
|
|
@@ -12915,7 +13273,7 @@
|
|
| 12915 |
"attributes": {}
|
| 12916 |
}
|
| 12917 |
},
|
| 12918 |
-
"total_flos": 9.
|
| 12919 |
"train_batch_size": 1,
|
| 12920 |
"trial_name": null,
|
| 12921 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.9550708746713874,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 18500,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 12896 |
"eval_samples_per_second": 2.929,
|
| 12897 |
"eval_steps_per_second": 2.929,
|
| 12898 |
"step": 18000
|
| 12899 |
+
},
|
| 12900 |
+
{
|
| 12901 |
+
"epoch": 1.9032854670594608,
|
| 12902 |
+
"grad_norm": 1.1796945333480835,
|
| 12903 |
+
"learning_rate": 1.4776056885979988e-05,
|
| 12904 |
+
"loss": 0.9963871955871582,
|
| 12905 |
+
"step": 18010
|
| 12906 |
+
},
|
| 12907 |
+
{
|
| 12908 |
+
"epoch": 1.9043423121127654,
|
| 12909 |
+
"grad_norm": 0.9647114872932434,
|
| 12910 |
+
"learning_rate": 1.4750799044087513e-05,
|
| 12911 |
+
"loss": 1.006678581237793,
|
| 12912 |
+
"step": 18020
|
| 12913 |
+
},
|
| 12914 |
+
{
|
| 12915 |
+
"epoch": 1.90539915716607,
|
| 12916 |
+
"grad_norm": 0.8601939678192139,
|
| 12917 |
+
"learning_rate": 1.4725553771221918e-05,
|
| 12918 |
+
"loss": 0.9009536743164063,
|
| 12919 |
+
"step": 18030
|
| 12920 |
+
},
|
| 12921 |
+
{
|
| 12922 |
+
"epoch": 1.9064560022193746,
|
| 12923 |
+
"grad_norm": 1.5132955312728882,
|
| 12924 |
+
"learning_rate": 1.4700321098342532e-05,
|
| 12925 |
+
"loss": 0.8412443161010742,
|
| 12926 |
+
"step": 18040
|
| 12927 |
+
},
|
| 12928 |
+
{
|
| 12929 |
+
"epoch": 1.9075128472726792,
|
| 12930 |
+
"grad_norm": 0.8298953175544739,
|
| 12931 |
+
"learning_rate": 1.4675101056393259e-05,
|
| 12932 |
+
"loss": 0.9263592720031738,
|
| 12933 |
+
"step": 18050
|
| 12934 |
+
},
|
| 12935 |
+
{
|
| 12936 |
+
"epoch": 1.9085696923259838,
|
| 12937 |
+
"grad_norm": 1.447222352027893,
|
| 12938 |
+
"learning_rate": 1.4649893676302484e-05,
|
| 12939 |
+
"loss": 0.9911669731140137,
|
| 12940 |
+
"step": 18060
|
| 12941 |
+
},
|
| 12942 |
+
{
|
| 12943 |
+
"epoch": 1.9096265373792884,
|
| 12944 |
+
"grad_norm": 1.3189270496368408,
|
| 12945 |
+
"learning_rate": 1.4624698988983093e-05,
|
| 12946 |
+
"loss": 0.9665737152099609,
|
| 12947 |
+
"step": 18070
|
| 12948 |
+
},
|
| 12949 |
+
{
|
| 12950 |
+
"epoch": 1.910683382432593,
|
| 12951 |
+
"grad_norm": 0.9956954121589661,
|
| 12952 |
+
"learning_rate": 1.4599517025332382e-05,
|
| 12953 |
+
"loss": 0.9582039833068847,
|
| 12954 |
+
"step": 18080
|
| 12955 |
+
},
|
| 12956 |
+
{
|
| 12957 |
+
"epoch": 1.9117402274858977,
|
| 12958 |
+
"grad_norm": 1.7711209058761597,
|
| 12959 |
+
"learning_rate": 1.4574347816232065e-05,
|
| 12960 |
+
"loss": 0.9661527633666992,
|
| 12961 |
+
"step": 18090
|
| 12962 |
+
},
|
| 12963 |
+
{
|
| 12964 |
+
"epoch": 1.9127970725392025,
|
| 12965 |
+
"grad_norm": 1.2257004976272583,
|
| 12966 |
+
"learning_rate": 1.4549191392548183e-05,
|
| 12967 |
+
"loss": 0.9142245292663574,
|
| 12968 |
+
"step": 18100
|
| 12969 |
+
},
|
| 12970 |
+
{
|
| 12971 |
+
"epoch": 1.913853917592507,
|
| 12972 |
+
"grad_norm": 0.9886197447776794,
|
| 12973 |
+
"learning_rate": 1.4524047785131153e-05,
|
| 12974 |
+
"loss": 0.9936203956604004,
|
| 12975 |
+
"step": 18110
|
| 12976 |
+
},
|
| 12977 |
+
{
|
| 12978 |
+
"epoch": 1.9149107626458117,
|
| 12979 |
+
"grad_norm": 1.3881423473358154,
|
| 12980 |
+
"learning_rate": 1.44989170248156e-05,
|
| 12981 |
+
"loss": 0.8859989166259765,
|
| 12982 |
+
"step": 18120
|
| 12983 |
+
},
|
| 12984 |
+
{
|
| 12985 |
+
"epoch": 1.9159676076991161,
|
| 12986 |
+
"grad_norm": 1.1195544004440308,
|
| 12987 |
+
"learning_rate": 1.4473799142420461e-05,
|
| 12988 |
+
"loss": 0.7937178134918212,
|
| 12989 |
+
"step": 18130
|
| 12990 |
+
},
|
| 12991 |
+
{
|
| 12992 |
+
"epoch": 1.917024452752421,
|
| 12993 |
+
"grad_norm": 1.0878417491912842,
|
| 12994 |
+
"learning_rate": 1.4448694168748832e-05,
|
| 12995 |
+
"loss": 0.9974874496459961,
|
| 12996 |
+
"step": 18140
|
| 12997 |
+
},
|
| 12998 |
+
{
|
| 12999 |
+
"epoch": 1.9180812978057253,
|
| 13000 |
+
"grad_norm": 1.1011698246002197,
|
| 13001 |
+
"learning_rate": 1.4423602134588037e-05,
|
| 13002 |
+
"loss": 0.9105242729187012,
|
| 13003 |
+
"step": 18150
|
| 13004 |
+
},
|
| 13005 |
+
{
|
| 13006 |
+
"epoch": 1.9191381428590302,
|
| 13007 |
+
"grad_norm": 1.4028321504592896,
|
| 13008 |
+
"learning_rate": 1.4398523070709447e-05,
|
| 13009 |
+
"loss": 0.9755278587341308,
|
| 13010 |
+
"step": 18160
|
| 13011 |
+
},
|
| 13012 |
+
{
|
| 13013 |
+
"epoch": 1.9201949879123346,
|
| 13014 |
+
"grad_norm": 1.2676739692687988,
|
| 13015 |
+
"learning_rate": 1.437345700786861e-05,
|
| 13016 |
+
"loss": 0.9609894752502441,
|
| 13017 |
+
"step": 18170
|
| 13018 |
+
},
|
| 13019 |
+
{
|
| 13020 |
+
"epoch": 1.9212518329656394,
|
| 13021 |
+
"grad_norm": 1.4375134706497192,
|
| 13022 |
+
"learning_rate": 1.4348403976805069e-05,
|
| 13023 |
+
"loss": 0.9367487907409668,
|
| 13024 |
+
"step": 18180
|
| 13025 |
+
},
|
| 13026 |
+
{
|
| 13027 |
+
"epoch": 1.9223086780189438,
|
| 13028 |
+
"grad_norm": 0.8992677330970764,
|
| 13029 |
+
"learning_rate": 1.4323364008242446e-05,
|
| 13030 |
+
"loss": 0.9339933395385742,
|
| 13031 |
+
"step": 18190
|
| 13032 |
+
},
|
| 13033 |
+
{
|
| 13034 |
+
"epoch": 1.9233655230722486,
|
| 13035 |
+
"grad_norm": 1.4033421277999878,
|
| 13036 |
+
"learning_rate": 1.4298337132888268e-05,
|
| 13037 |
+
"loss": 1.0861198425292968,
|
| 13038 |
+
"step": 18200
|
| 13039 |
+
},
|
| 13040 |
+
{
|
| 13041 |
+
"epoch": 1.9244223681255532,
|
| 13042 |
+
"grad_norm": 1.5736299753189087,
|
| 13043 |
+
"learning_rate": 1.4273323381434078e-05,
|
| 13044 |
+
"loss": 0.8618276596069336,
|
| 13045 |
+
"step": 18210
|
| 13046 |
+
},
|
| 13047 |
+
{
|
| 13048 |
+
"epoch": 1.9254792131788578,
|
| 13049 |
+
"grad_norm": 1.0226414203643799,
|
| 13050 |
+
"learning_rate": 1.4248322784555267e-05,
|
| 13051 |
+
"loss": 0.9526123046875,
|
| 13052 |
+
"step": 18220
|
| 13053 |
+
},
|
| 13054 |
+
{
|
| 13055 |
+
"epoch": 1.9265360582321625,
|
| 13056 |
+
"grad_norm": 1.0963102579116821,
|
| 13057 |
+
"learning_rate": 1.4223335372911158e-05,
|
| 13058 |
+
"loss": 0.8445890426635743,
|
| 13059 |
+
"step": 18230
|
| 13060 |
+
},
|
| 13061 |
+
{
|
| 13062 |
+
"epoch": 1.927592903285467,
|
| 13063 |
+
"grad_norm": 1.1241637468338013,
|
| 13064 |
+
"learning_rate": 1.4198361177144823e-05,
|
| 13065 |
+
"loss": 1.0155341148376464,
|
| 13066 |
+
"step": 18240
|
| 13067 |
+
},
|
| 13068 |
+
{
|
| 13069 |
+
"epoch": 1.9286497483387717,
|
| 13070 |
+
"grad_norm": 0.7999724745750427,
|
| 13071 |
+
"learning_rate": 1.4173400227883198e-05,
|
| 13072 |
+
"loss": 0.8864949226379395,
|
| 13073 |
+
"step": 18250
|
| 13074 |
+
},
|
| 13075 |
+
{
|
| 13076 |
+
"epoch": 1.9297065933920763,
|
| 13077 |
+
"grad_norm": 1.011460781097412,
|
| 13078 |
+
"learning_rate": 1.414845255573693e-05,
|
| 13079 |
+
"loss": 0.9812465667724609,
|
| 13080 |
+
"step": 18260
|
| 13081 |
+
},
|
| 13082 |
+
{
|
| 13083 |
+
"epoch": 1.930763438445381,
|
| 13084 |
+
"grad_norm": 0.9159804582595825,
|
| 13085 |
+
"learning_rate": 1.4123518191300428e-05,
|
| 13086 |
+
"loss": 0.8984330177307129,
|
| 13087 |
+
"step": 18270
|
| 13088 |
+
},
|
| 13089 |
+
{
|
| 13090 |
+
"epoch": 1.9318202834986855,
|
| 13091 |
+
"grad_norm": 1.0511664152145386,
|
| 13092 |
+
"learning_rate": 1.4098597165151717e-05,
|
| 13093 |
+
"loss": 0.9282947540283203,
|
| 13094 |
+
"step": 18280
|
| 13095 |
+
},
|
| 13096 |
+
{
|
| 13097 |
+
"epoch": 1.9328771285519901,
|
| 13098 |
+
"grad_norm": 1.3192466497421265,
|
| 13099 |
+
"learning_rate": 1.4073689507852527e-05,
|
| 13100 |
+
"loss": 1.0243149757385255,
|
| 13101 |
+
"step": 18290
|
| 13102 |
+
},
|
| 13103 |
+
{
|
| 13104 |
+
"epoch": 1.9339339736052947,
|
| 13105 |
+
"grad_norm": 1.1828497648239136,
|
| 13106 |
+
"learning_rate": 1.4048795249948148e-05,
|
| 13107 |
+
"loss": 0.9151687622070312,
|
| 13108 |
+
"step": 18300
|
| 13109 |
+
},
|
| 13110 |
+
{
|
| 13111 |
+
"epoch": 1.9349908186585996,
|
| 13112 |
+
"grad_norm": 1.1846466064453125,
|
| 13113 |
+
"learning_rate": 1.4023914421967487e-05,
|
| 13114 |
+
"loss": 0.9660979270935058,
|
| 13115 |
+
"step": 18310
|
| 13116 |
+
},
|
| 13117 |
+
{
|
| 13118 |
+
"epoch": 1.936047663711904,
|
| 13119 |
+
"grad_norm": 1.395215630531311,
|
| 13120 |
+
"learning_rate": 1.3999047054422915e-05,
|
| 13121 |
+
"loss": 0.9342232704162597,
|
| 13122 |
+
"step": 18320
|
| 13123 |
+
},
|
| 13124 |
+
{
|
| 13125 |
+
"epoch": 1.9371045087652088,
|
| 13126 |
+
"grad_norm": 1.2342299222946167,
|
| 13127 |
+
"learning_rate": 1.397419317781036e-05,
|
| 13128 |
+
"loss": 0.92709321975708,
|
| 13129 |
+
"step": 18330
|
| 13130 |
+
},
|
| 13131 |
+
{
|
| 13132 |
+
"epoch": 1.9381613538185132,
|
| 13133 |
+
"grad_norm": 1.1838783025741577,
|
| 13134 |
+
"learning_rate": 1.3949352822609158e-05,
|
| 13135 |
+
"loss": 0.9767477989196778,
|
| 13136 |
+
"step": 18340
|
| 13137 |
+
},
|
| 13138 |
+
{
|
| 13139 |
+
"epoch": 1.939218198871818,
|
| 13140 |
+
"grad_norm": 1.0344548225402832,
|
| 13141 |
+
"learning_rate": 1.3924526019282114e-05,
|
| 13142 |
+
"loss": 0.952883243560791,
|
| 13143 |
+
"step": 18350
|
| 13144 |
+
},
|
| 13145 |
+
{
|
| 13146 |
+
"epoch": 1.9402750439251224,
|
| 13147 |
+
"grad_norm": 1.2441740036010742,
|
| 13148 |
+
"learning_rate": 1.3899712798275338e-05,
|
| 13149 |
+
"loss": 0.9649330139160156,
|
| 13150 |
+
"step": 18360
|
| 13151 |
+
},
|
| 13152 |
+
{
|
| 13153 |
+
"epoch": 1.9413318889784272,
|
| 13154 |
+
"grad_norm": 1.2067946195602417,
|
| 13155 |
+
"learning_rate": 1.3874913190018363e-05,
|
| 13156 |
+
"loss": 1.1383502960205079,
|
| 13157 |
+
"step": 18370
|
| 13158 |
+
},
|
| 13159 |
+
{
|
| 13160 |
+
"epoch": 1.9423887340317316,
|
| 13161 |
+
"grad_norm": 1.0177148580551147,
|
| 13162 |
+
"learning_rate": 1.385012722492397e-05,
|
| 13163 |
+
"loss": 0.9787127494812011,
|
| 13164 |
+
"step": 18380
|
| 13165 |
+
},
|
| 13166 |
+
{
|
| 13167 |
+
"epoch": 1.9434455790850365,
|
| 13168 |
+
"grad_norm": 0.9557192325592041,
|
| 13169 |
+
"learning_rate": 1.3825354933388269e-05,
|
| 13170 |
+
"loss": 0.897181510925293,
|
| 13171 |
+
"step": 18390
|
| 13172 |
+
},
|
| 13173 |
+
{
|
| 13174 |
+
"epoch": 1.944502424138341,
|
| 13175 |
+
"grad_norm": 1.5747487545013428,
|
| 13176 |
+
"learning_rate": 1.3800596345790523e-05,
|
| 13177 |
+
"loss": 0.9553994178771973,
|
| 13178 |
+
"step": 18400
|
| 13179 |
+
},
|
| 13180 |
+
{
|
| 13181 |
+
"epoch": 1.9455592691916457,
|
| 13182 |
+
"grad_norm": 1.0466935634613037,
|
| 13183 |
+
"learning_rate": 1.3775851492493242e-05,
|
| 13184 |
+
"loss": 0.8787156105041504,
|
| 13185 |
+
"step": 18410
|
| 13186 |
+
},
|
| 13187 |
+
{
|
| 13188 |
+
"epoch": 1.9466161142449503,
|
| 13189 |
+
"grad_norm": 1.2227885723114014,
|
| 13190 |
+
"learning_rate": 1.37511204038421e-05,
|
| 13191 |
+
"loss": 0.946174144744873,
|
| 13192 |
+
"step": 18420
|
| 13193 |
+
},
|
| 13194 |
+
{
|
| 13195 |
+
"epoch": 1.947672959298255,
|
| 13196 |
+
"grad_norm": 1.0443825721740723,
|
| 13197 |
+
"learning_rate": 1.3726403110165853e-05,
|
| 13198 |
+
"loss": 1.0222416877746583,
|
| 13199 |
+
"step": 18430
|
| 13200 |
+
},
|
| 13201 |
+
{
|
| 13202 |
+
"epoch": 1.9487298043515595,
|
| 13203 |
+
"grad_norm": 1.4991309642791748,
|
| 13204 |
+
"learning_rate": 1.3701699641776373e-05,
|
| 13205 |
+
"loss": 0.8525497436523437,
|
| 13206 |
+
"step": 18440
|
| 13207 |
+
},
|
| 13208 |
+
{
|
| 13209 |
+
"epoch": 1.9497866494048641,
|
| 13210 |
+
"grad_norm": 1.3419233560562134,
|
| 13211 |
+
"learning_rate": 1.3677010028968546e-05,
|
| 13212 |
+
"loss": 0.9364970207214356,
|
| 13213 |
+
"step": 18450
|
| 13214 |
+
},
|
| 13215 |
+
{
|
| 13216 |
+
"epoch": 1.9508434944581687,
|
| 13217 |
+
"grad_norm": 1.253547191619873,
|
| 13218 |
+
"learning_rate": 1.3652334302020303e-05,
|
| 13219 |
+
"loss": 0.9723645210266113,
|
| 13220 |
+
"step": 18460
|
| 13221 |
+
},
|
| 13222 |
+
{
|
| 13223 |
+
"epoch": 1.9519003395114733,
|
| 13224 |
+
"grad_norm": 1.5022222995758057,
|
| 13225 |
+
"learning_rate": 1.3627672491192522e-05,
|
| 13226 |
+
"loss": 0.9641678810119629,
|
| 13227 |
+
"step": 18470
|
| 13228 |
+
},
|
| 13229 |
+
{
|
| 13230 |
+
"epoch": 1.952957184564778,
|
| 13231 |
+
"grad_norm": 1.2753891944885254,
|
| 13232 |
+
"learning_rate": 1.3603024626729021e-05,
|
| 13233 |
+
"loss": 0.9783175468444825,
|
| 13234 |
+
"step": 18480
|
| 13235 |
+
},
|
| 13236 |
+
{
|
| 13237 |
+
"epoch": 1.9540140296180826,
|
| 13238 |
+
"grad_norm": 1.638152837753296,
|
| 13239 |
+
"learning_rate": 1.35783907388565e-05,
|
| 13240 |
+
"loss": 0.956610107421875,
|
| 13241 |
+
"step": 18490
|
| 13242 |
+
},
|
| 13243 |
+
{
|
| 13244 |
+
"epoch": 1.9550708746713874,
|
| 13245 |
+
"grad_norm": 1.3578416109085083,
|
| 13246 |
+
"learning_rate": 1.3553770857784553e-05,
|
| 13247 |
+
"loss": 0.9685772895812989,
|
| 13248 |
+
"step": 18500
|
| 13249 |
+
},
|
| 13250 |
+
{
|
| 13251 |
+
"epoch": 1.9550708746713874,
|
| 13252 |
+
"eval_loss": 0.4817064702510834,
|
| 13253 |
+
"eval_runtime": 3413.9531,
|
| 13254 |
+
"eval_samples_per_second": 2.608,
|
| 13255 |
+
"eval_steps_per_second": 2.608,
|
| 13256 |
+
"step": 18500
|
| 13257 |
}
|
| 13258 |
],
|
| 13259 |
"logging_steps": 10,
|
|
|
|
| 13273 |
"attributes": {}
|
| 13274 |
}
|
| 13275 |
},
|
| 13276 |
+
"total_flos": 9.985879157251277e+17,
|
| 13277 |
"train_batch_size": 1,
|
| 13278 |
"trial_name": null,
|
| 13279 |
"trial_params": null
|