Training in progress, step 1875

adapter_config.json

@@ -1,5 +1,5 @@
 {
-  "base_model_name_or_path": "checkpoint_saves/Llama_13b/last-checkpoint",
+  "base_model_name_or_path": "checkpoint_saves/Llama_2_13b/checkpoint-145",
   "bias": "none",
   "fan_in_fan_out": false,
   "inference_mode": true,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:55936da28ccaa45e5ddca8dcdc226991b73479c6b56eb976c62b4594f394ab98
+oid sha256:3e5ac37515d11cff8718e9362a9ec10874b22f219b1bc73f65c9adc8a1e7e2d1
 size 104900720

last-checkpoint/README.md

@@ -5,5 +5,6 @@ library_name: peft
 
 ### Framework versions
 
+- PEFT 0.4.0.dev0
 
 - PEFT 0.4.0.dev0

last-checkpoint/adapter_config.json

@@ -1,5 +1,5 @@
 {
-  "base_model_name_or_path": "checkpoint_saves/Llama_13b/last-checkpoint",
+  "base_model_name_or_path": "checkpoint_saves/Llama_2_13b/checkpoint-145",
   "bias": "none",
   "fan_in_fan_out": false,
   "inference_mode": true,

last-checkpoint/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:55936da28ccaa45e5ddca8dcdc226991b73479c6b56eb976c62b4594f394ab98
+oid sha256:3e5ac37515d11cff8718e9362a9ec10874b22f219b1bc73f65c9adc8a1e7e2d1
 size 104900720

last-checkpoint/global_step1875/zero_pp_rank_0_mp_rank_00_model_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b86c8546dd1666401b84d90ca0b94ef6cb6c0b73117ae6c1d687ddb1e11f648b
+oid sha256:dc3aba48715eba0d62814c116be69cdd8db40c6bc42ce0caf09bc8fd54ff0bad
 size 6508524919

last-checkpoint/global_step1875/zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fee3a7fd29d3119d2b21bf40c50d75a2eebabc650ed9e0e16b2e8bcfe9449031
+oid sha256:482603883dc54fe0438c3e4d2df1653c84bf964b9ee4150b6df3ed7c6c65a0ad
 size 39324734

last-checkpoint/global_step1875/zero_pp_rank_1_mp_rank_00_model_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:86353a275442ecc1818f29ee23975deedfc0c42b7dc80def1b26bff124460652
+oid sha256:ac6cda5d42888c8bd34166fdf0673c46cc2cc6720eb9d95e7f52ae00c200eee4
 size 6508524919

last-checkpoint/global_step1875/zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bb9c8eabf344996ecedd94f084d70e33300335e97cdc132bc047b014185c6059
+oid sha256:2212e8fb1026d9c4a234a31b92fcbc3b9757e4661b361ba24c109fc438bd5efc
 size 39324734

last-checkpoint/global_step1875/zero_pp_rank_2_mp_rank_00_model_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:25055a4966ab33becc0bcb18d051dd85b05d08342ee8e7609d2649be7f0fe9d4
+oid sha256:7e0efab9f86a1fc4cdefe9209f280e916eb88e9f639304743a62c915be58130a
 size 6508524919

last-checkpoint/global_step1875/zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a1a1d5289fa7a41de645b8cb7890fb4ed7624b274892688cef21dfb5703cc831
+oid sha256:92c1d2000930eb4b557d099f2ddd740c2f3eb3cee1071cac00368cfc8d837f4d
 size 39324734

last-checkpoint/global_step1875/zero_pp_rank_3_mp_rank_00_model_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:da7d9d6595b20c78f8650a456533ca3ca7fd8a0d09c8b080efd314c39ab51fc7
+oid sha256:67f9ce15873371920fc37cb57c0c541b4ebfc5a589c2f344f8900182569870e4
 size 6508524919

last-checkpoint/global_step1875/zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9bef3cc71b28be9da43756109bd915e5a8693bc2e9f945c3a2c35f77d8fd5a36
+oid sha256:e658ab4cd3a5fb2a89ca0cc414ca8859577c5643624125d62142c5fb1e5ea419
 size 39324734

last-checkpoint/global_step1875/zero_pp_rank_4_mp_rank_00_model_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4dd45dd6271af5951105ff05aa28f48e34d0aa71ad8e9374c030e8495033d274
+oid sha256:bfbc572ac2614b0f71af3cece56fd993f587bc61dd630482ade88c179db97a9f
 size 6508524919

last-checkpoint/global_step1875/zero_pp_rank_4_mp_rank_00_optim_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c9c7d905697ce3ebcb3108c7a8d9c377f03e508805123a11d78736db0628e3b0
+oid sha256:0d1938abf09346d77d8fc8232f5b275addd5d5c5a3b0741ddb694c65cb509083
 size 39324734

last-checkpoint/global_step1875/zero_pp_rank_5_mp_rank_00_model_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7214390febf54a55dfed233c4252b5a5e1fe4a2e5a1dce4c268ff0b8af78add1
+oid sha256:56a74cc6cf0057fda63ded0033822c845711207b58de56e6571f53dc7626c1d5
 size 6508524919

last-checkpoint/global_step1875/zero_pp_rank_5_mp_rank_00_optim_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1d2e9a3e15eab72cb27e2ad212925c0bcc40c109704d5c12d76a00c783c0714b
+oid sha256:4135022cae2c7b096309646b76bc23b5b2be5ef498db8b11dad0f656f57cc589
 size 39324734

last-checkpoint/global_step1875/zero_pp_rank_6_mp_rank_00_model_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a4d315b42177b90d73673c0917af534743f494256bd881f5858dd280d4432bbb
+oid sha256:a07cf5614bb028d36f6496e22a69c00112170bb48a7fa5769c999304030ab0e1
 size 6508524919

last-checkpoint/global_step1875/zero_pp_rank_6_mp_rank_00_optim_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5bb2afbc9c7cb4f347271f6e4391eb9c7956b9e8fc68c0117bf80cb52c8cba7c
+oid sha256:25a86ac1a1fea9b8099833730b3bb980c521629a67af4f8ddfbe286dd38bfdb1
 size 39324734

last-checkpoint/global_step1875/zero_pp_rank_7_mp_rank_00_model_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f6a90eac5198209d06ad10d0085cb23471f4de6a307c6b2d74b5cb7067c96729
+oid sha256:acc28e5d4b9ee438f714f9ba46d05afb2425a8ac1bb7fa1fc700214d52cbd723
 size 6508524919

last-checkpoint/global_step1875/zero_pp_rank_7_mp_rank_00_optim_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b1e2129e9ba4b98ab3bf5f0c3afc2e58206d9d375a9163c28300e85e1b38fe30
+oid sha256:647e8061c41695b8f1c8597798fc9bb8112c1b108e26a0bc6483fb4754c5af8e
 size 39324734

last-checkpoint/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9d914bd28016f6122c43d9f2c83724d2eb3fb30e66e779af973bb2d54cc8392b
+oid sha256:285a047ead58adb79ee83ad0db3adbbda92d36a204c3cc5c668c1f063f1637c4
 size 21687

last-checkpoint/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6211fb1ee2fec138ceb8c47005fc9bc2418104c532ee8dfc1b97c6ff0a4d8ca6
+oid sha256:60722f7ef4638cf0aae77ec645716aada21f8d8687b56f379dbc90416a04543e
 size 21687

last-checkpoint/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5e6e93d0e06b92d456cdf9ae8d458dc04fe62bcaff12dcfd66e34ec4d96610a3
+oid sha256:29dcfebb85aa2f8ea2e4a051ecf97ef2dd4581dd56230e161c1cc07fbda1e938
 size 21687

last-checkpoint/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bd94ec78519765c3ee8dc3b4493f42b5fd0c4cd2976fd129e1f8ac18f99ae6d2
+oid sha256:0c0d3ac388816bd596a86ef1cba1c1f18e9bef020642f802d9bef6d67ab7a3b6
 size 21687

last-checkpoint/rng_state_4.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:63751f591f898abcc08b990d99973e8f008e12981693a109dce30ef0cd238781
+oid sha256:d55982f4a550feb1d76343d94cf85e3a96a0bafff49d034b23e3cf828350f12d
 size 21687

last-checkpoint/rng_state_5.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ee6e3fa3cef42dc11afb890036889f87f425dfc577cbf9e1d5de01bb444eb373
+oid sha256:f3472a83b969c6766805fbb6eac6e06e1fb0d1b413c8af2805c1a3f29eddf796
 size 21687

last-checkpoint/rng_state_6.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:61d84c5c8cd6aa95d17a38e63e5b602a1ef6cf5067ed65357c055b82af7e9406
+oid sha256:f99091da41f7eae5cc1767b133376cec72866d89d8f1312ec50333f9096a4b7a
 size 21687

last-checkpoint/rng_state_7.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1182410896d797ec61be8102c6e7744e694a891c867f4e04ed79c5737df19e2c
+oid sha256:1c9584f36f14255afedeab1ad3c2cbf4ae4ac6d7e38ed1a1a53e90e42ab75cba
 size 21687

last-checkpoint/special_tokens_map.json

@@ -0,0 +1,6 @@
+{
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "pad_token": "</s>",
+  "unk_token": "<unk>"
+}

last-checkpoint/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
+size 499723

last-checkpoint/tokenizer_config.json

@@ -0,0 +1,14 @@
+{
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "device_map": "cuda",
+  "eos_token": "</s>",
+  "max_length": 256,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": null,
+  "padding": "max_length",
+  "sp_model_kwargs": {},
+  "tokenizer_class": "LlamaTokenizer",
+  "truncation": "longest_first",
+  "unk_token": "<unk>"
+}

last-checkpoint/trainer_state.json

@@ -10,2257 +10,2257 @@
     {
       "epoch": 0.0,
       "learning_rate": 0.00013979400086720374,
-      "loss": 1.8118,
+      "loss": 0.0,
       "step": 5
     },
     {
       "epoch": 0.0,
       "learning_rate": 0.00019999999999999998,
-      "loss": 1.6525,
+      "loss": 0.0,
       "step": 10
     },
     {
       "epoch": 0.01,
       "learning_rate": 0.0001998460354118553,
-      "loss": 1.5208,
+      "loss": 0.0,
       "step": 15
     },
     {
       "epoch": 0.01,
       "learning_rate": 0.00019965357967667439,
-      "loss": 1.4179,
+      "loss": 0.0,
       "step": 20
     },
     {
       "epoch": 0.01,
       "learning_rate": 0.00019946112394149347,
-      "loss": 1.5179,
+      "loss": 0.0,
       "step": 25
     },
     {
       "epoch": 0.01,
       "learning_rate": 0.00019926866820631255,
-      "loss": 1.4164,
+      "loss": 0.0,
       "step": 30
     },
     {
       "epoch": 0.01,
       "learning_rate": 0.00019907621247113163,
-      "loss": 1.3599,
+      "loss": 0.0,
       "step": 35
     },
     {
       "epoch": 0.02,
       "learning_rate": 0.00019888375673595074,
-      "loss": 1.4631,
+      "loss": 0.0,
       "step": 40
     },
     {
       "epoch": 0.02,
       "learning_rate": 0.00019869130100076983,
-      "loss": 1.4569,
+      "loss": 0.0,
       "step": 45
     },
     {
       "epoch": 0.02,
       "learning_rate": 0.00019849884526558894,
-      "loss": 1.4491,
+      "loss": 0.0,
       "step": 50
     },
     {
       "epoch": 0.02,
       "learning_rate": 0.00019830638953040802,
-      "loss": 1.4852,
+      "loss": 0.0,
       "step": 55
     },
     {
       "epoch": 0.02,
       "learning_rate": 0.0001981139337952271,
-      "loss": 1.5018,
+      "loss": 0.0,
       "step": 60
     },
     {
       "epoch": 0.02,
       "learning_rate": 0.00019792147806004618,
-      "loss": 1.4777,
+      "loss": 0.0,
       "step": 65
     },
     {
       "epoch": 0.03,
       "learning_rate": 0.0001977290223248653,
-      "loss": 1.4531,
+      "loss": 0.0,
       "step": 70
     },
     {
       "epoch": 0.03,
       "learning_rate": 0.00019753656658968438,
-      "loss": 1.3477,
+      "loss": 0.0,
       "step": 75
     },
     {
       "epoch": 0.03,
       "learning_rate": 0.0001973441108545035,
-      "loss": 1.5621,
+      "loss": 0.0,
       "step": 80
     },
     {
       "epoch": 0.03,
       "learning_rate": 0.00019715165511932257,
-      "loss": 1.3823,
+      "loss": 0.0,
       "step": 85
     },
     {
       "epoch": 0.03,
       "learning_rate": 0.00019695919938414165,
-      "loss": 1.4521,
+      "loss": 0.0,
       "step": 90
     },
     {
       "epoch": 0.04,
       "learning_rate": 0.00019676674364896076,
-      "loss": 1.4822,
+      "loss": 0.0,
       "step": 95
     },
     {
       "epoch": 0.04,
       "learning_rate": 0.00019657428791377982,
-      "loss": 1.4313,
+      "loss": 0.0,
       "step": 100
     },
     {
       "epoch": 0.04,
       "learning_rate": 0.00019638183217859893,
-      "loss": 1.3581,
+      "loss": 0.0,
       "step": 105
     },
     {
       "epoch": 0.04,
       "learning_rate": 0.000196189376443418,
-      "loss": 1.3917,
+      "loss": 0.0,
       "step": 110
     },
     {
       "epoch": 0.04,
       "learning_rate": 0.00019599692070823712,
-      "loss": 1.4632,
+      "loss": 0.0,
       "step": 115
     },
     {
       "epoch": 0.05,
       "learning_rate": 0.0001958044649730562,
-      "loss": 1.4535,
+      "loss": 0.0,
       "step": 120
     },
     {
       "epoch": 0.05,
       "learning_rate": 0.0001956120092378753,
-      "loss": 1.487,
+      "loss": 0.0,
       "step": 125
     },
     {
       "epoch": 0.05,
       "learning_rate": 0.0001954195535026944,
-      "loss": 1.3851,
+      "loss": 0.0,
       "step": 130
     },
     {
       "epoch": 0.05,
       "learning_rate": 0.00019522709776751348,
-      "loss": 1.3467,
+      "loss": 0.0,
       "step": 135
     },
     {
       "epoch": 0.05,
       "learning_rate": 0.00019503464203233256,
-      "loss": 1.4262,
+      "loss": 0.0,
       "step": 140
     },
     {
       "epoch": 0.06,
       "learning_rate": 0.00019484218629715167,
-      "loss": 1.326,
+      "loss": 0.0,
       "step": 145
     },
     {
       "epoch": 0.06,
       "learning_rate": 0.00019464973056197075,
-      "loss": 1.471,
+      "loss": 0.0,
       "step": 150
     },
     {
       "epoch": 0.06,
       "learning_rate": 0.00019445727482678984,
-      "loss": 1.4928,
+      "loss": 0.0,
       "step": 155
     },
     {
       "epoch": 0.06,
       "learning_rate": 0.00019426481909160895,
-      "loss": 1.3059,
+      "loss": 0.0,
       "step": 160
     },
     {
       "epoch": 0.06,
       "learning_rate": 0.00019407236335642803,
-      "loss": 1.3448,
+      "loss": 0.0,
       "step": 165
     },
     {
       "epoch": 0.07,
       "learning_rate": 0.0001938799076212471,
-      "loss": 1.4605,
+      "loss": 0.0,
       "step": 170
     },
     {
       "epoch": 0.07,
       "learning_rate": 0.0001936874518860662,
-      "loss": 1.3755,
+      "loss": 0.0,
       "step": 175
     },
     {
       "epoch": 0.07,
       "learning_rate": 0.0001934949961508853,
-      "loss": 1.4118,
+      "loss": 0.0,
       "step": 180
     },
     {
       "epoch": 0.07,
       "learning_rate": 0.0001933025404157044,
-      "loss": 1.302,
+      "loss": 0.0,
       "step": 185
     },
     {
       "epoch": 0.07,
       "learning_rate": 0.0001931100846805235,
-      "loss": 1.441,
+      "loss": 0.0,
       "step": 190
     },
     {
       "epoch": 0.07,
       "learning_rate": 0.00019291762894534258,
-      "loss": 1.4891,
+      "loss": 0.0,
       "step": 195
     },
     {
       "epoch": 0.08,
       "learning_rate": 0.0001927251732101617,
-      "loss": 1.4324,
+      "loss": 0.0,
       "step": 200
     },
     {
       "epoch": 0.08,
       "learning_rate": 0.00019253271747498077,
-      "loss": 1.3829,
+      "loss": 0.0,
       "step": 205
     },
     {
       "epoch": 0.08,
       "learning_rate": 0.00019234026173979986,
-      "loss": 1.3732,
+      "loss": 0.0,
       "step": 210
     },
     {
       "epoch": 0.08,
       "learning_rate": 0.00019214780600461894,
-      "loss": 1.3262,
+      "loss": 0.0,
       "step": 215
     },
     {
       "epoch": 0.08,
       "learning_rate": 0.00019195535026943802,
-      "loss": 1.3651,
+      "loss": 0.0,
       "step": 220
     },
     {
       "epoch": 0.09,
       "learning_rate": 0.00019176289453425713,
-      "loss": 1.4975,
+      "loss": 0.0,
       "step": 225
     },
     {
       "epoch": 0.09,
       "learning_rate": 0.00019157043879907621,
-      "loss": 1.341,
+      "loss": 0.0,
       "step": 230
     },
     {
       "epoch": 0.09,
       "learning_rate": 0.00019137798306389532,
-      "loss": 1.4237,
+      "loss": 0.0,
       "step": 235
     },
     {
       "epoch": 0.09,
       "learning_rate": 0.0001911855273287144,
-      "loss": 1.444,
+      "loss": 0.0,
       "step": 240
     },
     {
       "epoch": 0.09,
       "learning_rate": 0.0001909930715935335,
-      "loss": 1.4071,
+      "loss": 0.0,
       "step": 245
     },
     {
       "epoch": 0.1,
       "learning_rate": 0.00019080061585835257,
-      "loss": 1.3715,
+      "loss": 0.0,
       "step": 250
     },
     {
       "epoch": 0.1,
       "learning_rate": 0.00019060816012317168,
-      "loss": 1.3385,
+      "loss": 0.0,
       "step": 255
     },
     {
       "epoch": 0.1,
       "learning_rate": 0.00019041570438799076,
-      "loss": 1.4171,
+      "loss": 0.0,
       "step": 260
     },
     {
       "epoch": 0.1,
       "learning_rate": 0.00019022324865280987,
-      "loss": 1.3596,
+      "loss": 0.0,
       "step": 265
     },
     {
       "epoch": 0.1,
       "learning_rate": 0.00019003079291762896,
-      "loss": 1.3444,
+      "loss": 0.0,
       "step": 270
     },
     {
       "epoch": 0.11,
       "learning_rate": 0.00018983833718244807,
-      "loss": 1.4197,
+      "loss": 0.0,
       "step": 275
     },
     {
       "epoch": 0.11,
       "learning_rate": 0.00018964588144726715,
-      "loss": 1.3405,
+      "loss": 0.0,
       "step": 280
     },
     {
       "epoch": 0.11,
       "learning_rate": 0.0001894534257120862,
-      "loss": 1.4371,
+      "loss": 0.0,
       "step": 285
     },
     {
       "epoch": 0.11,
       "learning_rate": 0.00018926096997690532,
-      "loss": 1.3825,
+      "loss": 0.0,
       "step": 290
     },
     {
       "epoch": 0.11,
       "learning_rate": 0.0001890685142417244,
-      "loss": 1.4089,
+      "loss": 0.0,
       "step": 295
     },
     {
       "epoch": 0.12,
       "learning_rate": 0.0001888760585065435,
-      "loss": 1.4502,
+      "loss": 0.0,
       "step": 300
     },
     {
       "epoch": 0.12,
       "learning_rate": 0.0001886836027713626,
-      "loss": 1.3097,
+      "loss": 0.0,
       "step": 305
     },
     {
       "epoch": 0.12,
       "learning_rate": 0.0001884911470361817,
-      "loss": 1.3422,
+      "loss": 0.0,
       "step": 310
     },
     {
       "epoch": 0.12,
       "learning_rate": 0.00018829869130100078,
-      "loss": 1.4088,
+      "loss": 0.0,
       "step": 315
     },
     {
       "epoch": 0.12,
       "learning_rate": 0.00018810623556581987,
-      "loss": 1.3372,
+      "loss": 0.0,
       "step": 320
     },
     {
       "epoch": 0.12,
       "learning_rate": 0.00018791377983063895,
-      "loss": 1.3307,
+      "loss": 0.0,
       "step": 325
     },
     {
       "epoch": 0.13,
       "learning_rate": 0.00018772132409545806,
-      "loss": 1.4299,
+      "loss": 0.0,
       "step": 330
     },
     {
       "epoch": 0.13,
       "learning_rate": 0.00018752886836027714,
-      "loss": 1.4264,
+      "loss": 0.0,
       "step": 335
     },
     {
       "epoch": 0.13,
       "learning_rate": 0.00018733641262509625,
-      "loss": 1.3151,
+      "loss": 0.0,
       "step": 340
     },
     {
       "epoch": 0.13,
       "learning_rate": 0.00018714395688991533,
-      "loss": 1.3675,
+      "loss": 0.0,
       "step": 345
     },
     {
       "epoch": 0.13,
       "learning_rate": 0.00018695150115473442,
-      "loss": 1.3794,
+      "loss": 0.0,
       "step": 350
     },
     {
       "epoch": 0.14,
       "learning_rate": 0.00018675904541955353,
-      "loss": 1.3892,
+      "loss": 0.0,
       "step": 355
     },
     {
       "epoch": 0.14,
       "learning_rate": 0.00018656658968437258,
-      "loss": 1.4612,
+      "loss": 0.0,
       "step": 360
     },
     {
       "epoch": 0.14,
       "learning_rate": 0.0001863741339491917,
-      "loss": 1.283,
+      "loss": 0.0,
       "step": 365
     },
     {
       "epoch": 0.14,
       "learning_rate": 0.00018618167821401078,
-      "loss": 1.4235,
+      "loss": 0.0,
       "step": 370
     },
     {
       "epoch": 0.14,
       "learning_rate": 0.00018598922247882988,
-      "loss": 1.3036,
+      "loss": 0.0,
       "step": 375
     },
     {
       "epoch": 0.15,
       "learning_rate": 0.00018579676674364897,
-      "loss": 1.3886,
+      "loss": 0.0,
       "step": 380
     },
     {
       "epoch": 0.15,
       "learning_rate": 0.00018560431100846808,
-      "loss": 1.3226,
+      "loss": 0.0,
       "step": 385
     },
     {
       "epoch": 0.15,
       "learning_rate": 0.00018541185527328716,
-      "loss": 1.3966,
+      "loss": 0.0,
       "step": 390
     },
     {
       "epoch": 0.15,
       "learning_rate": 0.00018521939953810624,
-      "loss": 1.3835,
+      "loss": 0.0,
       "step": 395
     },
     {
       "epoch": 0.15,
       "learning_rate": 0.00018502694380292533,
-      "loss": 1.3544,
+      "loss": 0.0,
       "step": 400
     },
     {
       "epoch": 0.16,
       "learning_rate": 0.00018483448806774444,
-      "loss": 1.5449,
+      "loss": 0.0,
       "step": 405
     },
     {
       "epoch": 0.16,
       "learning_rate": 0.00018464203233256352,
-      "loss": 1.3814,
+      "loss": 0.0,
       "step": 410
     },
     {
       "epoch": 0.16,
       "learning_rate": 0.0001844495765973826,
-      "loss": 1.4272,
+      "loss": 0.0,
       "step": 415
     },
     {
       "epoch": 0.16,
       "learning_rate": 0.0001842571208622017,
-      "loss": 1.4256,
+      "loss": 0.0,
       "step": 420
     },
     {
       "epoch": 0.16,
       "learning_rate": 0.0001840646651270208,
-      "loss": 1.367,
+      "loss": 0.0,
       "step": 425
     },
     {
       "epoch": 0.17,
       "learning_rate": 0.00018387220939183988,
-      "loss": 1.3814,
+      "loss": 0.0,
       "step": 430
     },
     {
       "epoch": 0.17,
       "learning_rate": 0.00018367975365665896,
-      "loss": 1.351,
+      "loss": 0.0,
       "step": 435
     },
     {
       "epoch": 0.17,
       "learning_rate": 0.00018348729792147807,
-      "loss": 1.3966,
+      "loss": 0.0,
       "step": 440
     },
     {
       "epoch": 0.17,
       "learning_rate": 0.00018329484218629715,
-      "loss": 1.4355,
+      "loss": 0.0,
       "step": 445
     },
     {
       "epoch": 0.17,
       "learning_rate": 0.00018310238645111626,
-      "loss": 1.3635,
+      "loss": 0.0,
       "step": 450
     },
     {
       "epoch": 0.17,
       "learning_rate": 0.00018290993071593534,
-      "loss": 1.4373,
+      "loss": 0.0,
       "step": 455
     },
     {
       "epoch": 0.18,
       "learning_rate": 0.00018271747498075445,
-      "loss": 1.424,
+      "loss": 0.0,
       "step": 460
     },
     {
       "epoch": 0.18,
       "learning_rate": 0.00018252501924557354,
-      "loss": 1.4445,
+      "loss": 0.0,
       "step": 465
     },
     {
       "epoch": 0.18,
       "learning_rate": 0.00018233256351039262,
-      "loss": 1.4569,
+      "loss": 0.0,
       "step": 470
     },
     {
       "epoch": 0.18,
       "learning_rate": 0.0001821401077752117,
-      "loss": 1.2821,
+      "loss": 0.0,
       "step": 475
     },
     {
       "epoch": 0.18,
       "learning_rate": 0.00018194765204003079,
-      "loss": 1.4447,
+      "loss": 0.0,
       "step": 480
     },
     {
       "epoch": 0.19,
       "learning_rate": 0.0001817551963048499,
-      "loss": 1.5219,
+      "loss": 0.0,
       "step": 485
     },
     {
       "epoch": 0.19,
       "learning_rate": 0.00018156274056966898,
-      "loss": 1.3272,
+      "loss": 0.0,
       "step": 490
     },
     {
       "epoch": 0.19,
       "learning_rate": 0.0001813702848344881,
-      "loss": 1.3045,
+      "loss": 0.0,
       "step": 495
     },
     {
       "epoch": 0.19,
       "learning_rate": 0.00018117782909930717,
-      "loss": 1.4005,
+      "loss": 0.0,
       "step": 500
     },
     {
       "epoch": 0.19,
       "learning_rate": 0.00018098537336412625,
-      "loss": 1.277,
+      "loss": 0.0,
       "step": 505
     },
     {
       "epoch": 0.2,
       "learning_rate": 0.00018079291762894534,
-      "loss": 1.4171,
+      "loss": 0.0,
       "step": 510
     },
     {
       "epoch": 0.2,
       "learning_rate": 0.00018060046189376445,
-      "loss": 1.3951,
+      "loss": 0.0,
       "step": 515
     },
     {
       "epoch": 0.2,
       "learning_rate": 0.00018040800615858353,
-      "loss": 1.3163,
+      "loss": 0.0,
       "step": 520
     },
     {
       "epoch": 0.2,
       "learning_rate": 0.00018021555042340264,
-      "loss": 1.4673,
+      "loss": 0.0,
       "step": 525
     },
     {
       "epoch": 0.2,
       "learning_rate": 0.00018002309468822172,
-      "loss": 1.356,
+      "loss": 0.0,
       "step": 530
     },
     {
       "epoch": 0.21,
       "learning_rate": 0.00017983063895304083,
-      "loss": 1.4882,
+      "loss": 0.0,
       "step": 535
     },
     {
       "epoch": 0.21,
       "learning_rate": 0.00017963818321785991,
-      "loss": 1.4118,
+      "loss": 0.0,
       "step": 540
     },
     {
       "epoch": 0.21,
       "learning_rate": 0.00017944572748267897,
-      "loss": 1.4022,
+      "loss": 0.0,
       "step": 545
     },
     {
       "epoch": 0.21,
       "learning_rate": 0.00017925327174749808,
-      "loss": 1.4223,
+      "loss": 0.0,
       "step": 550
     },
     {
       "epoch": 0.21,
       "learning_rate": 0.00017906081601231716,
-      "loss": 1.3543,
+      "loss": 0.0,
       "step": 555
     },
     {
       "epoch": 0.22,
       "learning_rate": 0.00017886836027713627,
-      "loss": 1.2972,
+      "loss": 0.0,
       "step": 560
     },
     {
       "epoch": 0.22,
       "learning_rate": 0.00017867590454195535,
-      "loss": 1.3908,
+      "loss": 0.0,
       "step": 565
     },
     {
       "epoch": 0.22,
       "learning_rate": 0.00017848344880677446,
-      "loss": 1.4811,
+      "loss": 0.0,
       "step": 570
     },
     {
       "epoch": 0.22,
       "learning_rate": 0.00017829099307159355,
-      "loss": 1.4248,
+      "loss": 0.0,
       "step": 575
     },
     {
       "epoch": 0.22,
       "learning_rate": 0.00017809853733641263,
-      "loss": 1.4548,
+      "loss": 0.0,
       "step": 580
     },
     {
       "epoch": 0.22,
       "learning_rate": 0.0001779060816012317,
-      "loss": 1.4073,
+      "loss": 0.0,
       "step": 585
     },
     {
       "epoch": 0.23,
       "learning_rate": 0.00017771362586605082,
-      "loss": 1.3499,
+      "loss": 0.0,
       "step": 590
     },
     {
       "epoch": 0.23,
       "learning_rate": 0.0001775211701308699,
-      "loss": 1.2826,
+      "loss": 0.0,
       "step": 595
     },
     {
       "epoch": 0.23,
       "learning_rate": 0.00017732871439568902,
-      "loss": 1.4861,
+      "loss": 0.0,
       "step": 600
     },
     {
       "epoch": 0.23,
       "learning_rate": 0.0001771362586605081,
-      "loss": 1.4161,
+      "loss": 0.0,
       "step": 605
     },
     {
       "epoch": 0.23,
       "learning_rate": 0.00017694380292532718,
-      "loss": 1.3263,
+      "loss": 0.0,
       "step": 610
     },
     {
       "epoch": 0.24,
       "learning_rate": 0.00017675134719014626,
-      "loss": 1.3883,
+      "loss": 0.0,
       "step": 615
     },
     {
       "epoch": 0.24,
       "learning_rate": 0.00017655889145496535,
-      "loss": 1.3328,
+      "loss": 0.0,
       "step": 620
     },
     {
       "epoch": 0.24,
       "learning_rate": 0.00017636643571978446,
-      "loss": 1.3174,
+      "loss": 0.0,
       "step": 625
     },
     {
       "epoch": 0.24,
       "learning_rate": 0.00017617397998460354,
-      "loss": 1.3876,
+      "loss": 0.0,
       "step": 630
     },
     {
       "epoch": 0.24,
       "learning_rate": 0.00017598152424942265,
-      "loss": 1.3541,
+      "loss": 0.0,
       "step": 635
     },
     {
       "epoch": 0.25,
       "learning_rate": 0.00017578906851424173,
-      "loss": 1.3319,
+      "loss": 0.0,
       "step": 640
     },
     {
       "epoch": 0.25,
       "learning_rate": 0.00017559661277906084,
-      "loss": 1.2886,
+      "loss": 0.0,
       "step": 645
     },
     {
       "epoch": 0.25,
       "learning_rate": 0.00017540415704387992,
-      "loss": 1.3338,
+      "loss": 0.0,
       "step": 650
     },
     {
       "epoch": 0.25,
       "learning_rate": 0.000175211701308699,
-      "loss": 1.4036,
+      "loss": 0.0,
       "step": 655
     },
     {
       "epoch": 0.25,
       "learning_rate": 0.0001750192455735181,
-      "loss": 1.4543,
+      "loss": 0.0,
       "step": 660
     },
     {
       "epoch": 0.26,
       "learning_rate": 0.0001748267898383372,
-      "loss": 1.2902,
+      "loss": 0.0,
       "step": 665
     },
     {
       "epoch": 0.26,
       "learning_rate": 0.00017463433410315628,
-      "loss": 1.3255,
+      "loss": 0.0,
       "step": 670
     },
     {
       "epoch": 0.26,
       "learning_rate": 0.00017444187836797537,
-      "loss": 1.2875,
+      "loss": 0.0,
       "step": 675
     },
     {
       "epoch": 0.26,
       "learning_rate": 0.00017424942263279448,
-      "loss": 1.3105,
+      "loss": 0.0,
       "step": 680
     },
     {
       "epoch": 0.26,
       "learning_rate": 0.00017405696689761356,
-      "loss": 1.2533,
+      "loss": 0.0,
       "step": 685
     },
     {
       "epoch": 0.27,
       "learning_rate": 0.00017386451116243264,
-      "loss": 1.3827,
+      "loss": 0.0,
       "step": 690
     },
     {
       "epoch": 0.27,
       "learning_rate": 0.00017367205542725172,
-      "loss": 1.3549,
+      "loss": 0.0,
       "step": 695
     },
     {
       "epoch": 0.27,
       "learning_rate": 0.00017347959969207083,
-      "loss": 1.4396,
+      "loss": 0.0,
       "step": 700
     },
     {
       "epoch": 0.27,
       "learning_rate": 0.00017328714395688992,
-      "loss": 1.3179,
+      "loss": 0.0,
       "step": 705
     },
     {
       "epoch": 0.27,
       "learning_rate": 0.00017309468822170903,
-      "loss": 1.356,
+      "loss": 0.0,
       "step": 710
     },
     {
       "epoch": 0.27,
       "learning_rate": 0.0001729022324865281,
-      "loss": 1.2499,
+      "loss": 0.0,
       "step": 715
     },
     {
       "epoch": 0.28,
       "learning_rate": 0.00017270977675134722,
-      "loss": 1.4139,
+      "loss": 0.0,
       "step": 720
     },
     {
       "epoch": 0.28,
       "learning_rate": 0.0001725173210161663,
-      "loss": 1.249,
+      "loss": 0.0,
       "step": 725
     },
     {
       "epoch": 0.28,
       "learning_rate": 0.00017232486528098538,
-      "loss": 1.3054,
+      "loss": 0.0,
       "step": 730
     },
     {
       "epoch": 0.28,
       "learning_rate": 0.00017213240954580447,
-      "loss": 1.3415,
+      "loss": 0.0,
       "step": 735
     },
     {
       "epoch": 0.28,
       "learning_rate": 0.00017193995381062355,
-      "loss": 1.3323,
+      "loss": 0.0,
       "step": 740
     },
     {
       "epoch": 0.29,
       "learning_rate": 0.00017174749807544266,
-      "loss": 1.3559,
+      "loss": 0.0,
       "step": 745
     },
     {
       "epoch": 0.29,
       "learning_rate": 0.00017155504234026174,
-      "loss": 1.3771,
+      "loss": 0.0,
       "step": 750
     },
     {
       "epoch": 0.29,
       "learning_rate": 0.00017136258660508085,
-      "loss": 1.3811,
+      "loss": 0.0,
       "step": 755
     },
     {
       "epoch": 0.29,
       "learning_rate": 0.00017117013086989993,
-      "loss": 1.3644,
+      "loss": 0.0,
       "step": 760
     },
     {
       "epoch": 0.29,
       "learning_rate": 0.00017097767513471902,
-      "loss": 1.2619,
+      "loss": 0.0,
       "step": 765
     },
     {
       "epoch": 0.3,
       "learning_rate": 0.0001707852193995381,
-      "loss": 1.3795,
+      "loss": 0.0,
       "step": 770
     },
     {
       "epoch": 0.3,
       "learning_rate": 0.0001705927636643572,
-      "loss": 1.4482,
+      "loss": 0.0,
       "step": 775
     },
     {
       "epoch": 0.3,
       "learning_rate": 0.0001704003079291763,
-      "loss": 1.3213,
+      "loss": 0.0,
       "step": 780
     },
     {
       "epoch": 0.3,
       "learning_rate": 0.0001702078521939954,
-      "loss": 1.3406,
+      "loss": 0.0,
       "step": 785
     },
     {
       "epoch": 0.3,
       "learning_rate": 0.00017001539645881449,
-      "loss": 1.254,
+      "loss": 0.0,
       "step": 790
     },
     {
       "epoch": 0.31,
       "learning_rate": 0.0001698229407236336,
-      "loss": 1.4489,
+      "loss": 0.0,
       "step": 795
     },
     {
       "epoch": 0.31,
       "learning_rate": 0.00016963048498845268,
-      "loss": 1.3327,
+      "loss": 0.0,
       "step": 800
     },
     {
       "epoch": 0.31,
       "learning_rate": 0.00016943802925327173,
-      "loss": 1.2697,
+      "loss": 0.0,
       "step": 805
     },
     {
       "epoch": 0.31,
       "learning_rate": 0.00016924557351809084,
-      "loss": 1.3808,
+      "loss": 0.0,
       "step": 810
     },
     {
       "epoch": 0.31,
       "learning_rate": 0.00016905311778290993,
-      "loss": 1.3519,
+      "loss": 0.0,
       "step": 815
     },
     {
       "epoch": 0.32,
       "learning_rate": 0.00016886066204772904,
-      "loss": 1.3051,
+      "loss": 0.0,
       "step": 820
     },
     {
       "epoch": 0.32,
       "learning_rate": 0.00016866820631254812,
-      "loss": 1.3251,
+      "loss": 0.0,
       "step": 825
     },
     {
       "epoch": 0.32,
       "learning_rate": 0.00016847575057736723,
-      "loss": 1.3595,
+      "loss": 0.0,
       "step": 830
     },
     {
       "epoch": 0.32,
       "learning_rate": 0.0001682832948421863,
-      "loss": 1.366,
+      "loss": 0.0,
       "step": 835
     },
     {
       "epoch": 0.32,
       "learning_rate": 0.0001680908391070054,
-      "loss": 1.3157,
+      "loss": 0.0,
       "step": 840
     },
     {
       "epoch": 0.32,
       "learning_rate": 0.00016789838337182448,
-      "loss": 1.3809,
+      "loss": 0.0,
       "step": 845
     },
     {
       "epoch": 0.33,
       "learning_rate": 0.0001677059276366436,
-      "loss": 1.3305,
+      "loss": 0.0,
       "step": 850
     },
     {
       "epoch": 0.33,
       "learning_rate": 0.00016751347190146267,
-      "loss": 1.3454,
+      "loss": 0.0,
       "step": 855
     },
     {
       "epoch": 0.33,
       "learning_rate": 0.00016732101616628175,
-      "loss": 1.3914,
+      "loss": 0.0,
       "step": 860
     },
     {
       "epoch": 0.33,
       "learning_rate": 0.00016712856043110086,
-      "loss": 1.2801,
+      "loss": 0.0,
       "step": 865
     },
     {
       "epoch": 0.33,
       "learning_rate": 0.00016693610469591995,
-      "loss": 1.2451,
+      "loss": 0.0,
       "step": 870
     },
     {
       "epoch": 0.34,
       "learning_rate": 0.00016674364896073903,
-      "loss": 1.3802,
+      "loss": 0.0,
       "step": 875
     },
     {
       "epoch": 0.34,
       "learning_rate": 0.0001665511932255581,
-      "loss": 1.383,
+      "loss": 0.0,
       "step": 880
     },
     {
       "epoch": 0.34,
       "learning_rate": 0.00016635873749037722,
-      "loss": 1.3572,
+      "loss": 0.0,
       "step": 885
     },
     {
       "epoch": 0.34,
       "learning_rate": 0.0001661662817551963,
-      "loss": 1.381,
+      "loss": 0.0,
       "step": 890
     },
     {
       "epoch": 0.34,
       "learning_rate": 0.0001659738260200154,
-      "loss": 1.253,
+      "loss": 0.0,
       "step": 895
     },
     {
       "epoch": 0.35,
       "learning_rate": 0.0001657813702848345,
-      "loss": 1.3824,
+      "loss": 0.0,
       "step": 900
     },
     {
       "epoch": 0.35,
       "learning_rate": 0.0001655889145496536,
-      "loss": 1.3366,
+      "loss": 0.0,
       "step": 905
     },
     {
       "epoch": 0.35,
       "learning_rate": 0.0001653964588144727,
-      "loss": 1.2716,
+      "loss": 0.0,
       "step": 910
     },
     {
       "epoch": 0.35,
       "learning_rate": 0.00016520400307929177,
-      "loss": 1.317,
+      "loss": 0.0,
       "step": 915
     },
     {
       "epoch": 0.35,
       "learning_rate": 0.00016501154734411085,
-      "loss": 1.27,
+      "loss": 0.0,
       "step": 920
     },
     {
       "epoch": 0.36,
       "learning_rate": 0.00016481909160892994,
-      "loss": 1.463,
+      "loss": 0.0,
       "step": 925
     },
     {
       "epoch": 0.36,
       "learning_rate": 0.00016462663587374905,
-      "loss": 1.3101,
+      "loss": 0.0,
       "step": 930
     },
     {
       "epoch": 0.36,
       "learning_rate": 0.00016443418013856813,
-      "loss": 1.3305,
+      "loss": 0.0,
       "step": 935
     },
     {
       "epoch": 0.36,
       "learning_rate": 0.00016424172440338724,
-      "loss": 1.2637,
+      "loss": 0.0,
       "step": 940
     },
     {
       "epoch": 0.36,
       "learning_rate": 0.00016404926866820632,
-      "loss": 1.3564,
+      "loss": 0.0,
       "step": 945
     },
     {
       "epoch": 0.36,
       "learning_rate": 0.0001638568129330254,
-      "loss": 1.1473,
+      "loss": 0.0,
       "step": 950
     },
     {
       "epoch": 0.37,
       "learning_rate": 0.0001636643571978445,
-      "loss": 1.2369,
+      "loss": 0.0,
       "step": 955
     },
     {
       "epoch": 0.37,
       "learning_rate": 0.0001634719014626636,
-      "loss": 1.2508,
+      "loss": 0.0,
       "step": 960
     },
     {
       "epoch": 0.37,
       "learning_rate": 0.00016327944572748268,
-      "loss": 1.3019,
+      "loss": 0.0,
       "step": 965
     },
     {
       "epoch": 0.37,
       "learning_rate": 0.0001630869899923018,
-      "loss": 1.2893,
+      "loss": 0.0,
       "step": 970
     },
     {
       "epoch": 0.37,
       "learning_rate": 0.00016289453425712087,
-      "loss": 1.3052,
+      "loss": 0.0,
       "step": 975
     },
     {
       "epoch": 0.38,
       "learning_rate": 0.00016270207852193998,
-      "loss": 1.3312,
+      "loss": 0.0,
       "step": 980
     },
     {
       "epoch": 0.38,
       "learning_rate": 0.00016250962278675907,
-      "loss": 1.3215,
+      "loss": 0.0,
       "step": 985
     },
     {
       "epoch": 0.38,
       "learning_rate": 0.00016231716705157812,
-      "loss": 1.3382,
+      "loss": 0.0,
       "step": 990
     },
     {
       "epoch": 0.38,
       "learning_rate": 0.00016212471131639723,
-      "loss": 1.3263,
+      "loss": 0.0,
       "step": 995
     },
     {
       "epoch": 0.38,
       "learning_rate": 0.00016193225558121631,
-      "loss": 1.2434,
+      "loss": 0.0,
       "step": 1000
     },
     {
       "epoch": 0.39,
       "learning_rate": 0.00016173979984603542,
-      "loss": 1.3063,
+      "loss": 0.0,
       "step": 1005
     },
     {
       "epoch": 0.39,
       "learning_rate": 0.0001615473441108545,
-      "loss": 1.3254,
+      "loss": 0.0,
       "step": 1010
     },
     {
       "epoch": 0.39,
       "learning_rate": 0.00016135488837567362,
-      "loss": 1.2654,
+      "loss": 0.0,
       "step": 1015
     },
     {
       "epoch": 0.39,
       "learning_rate": 0.0001611624326404927,
-      "loss": 1.348,
+      "loss": 0.0,
       "step": 1020
     },
     {
       "epoch": 0.39,
       "learning_rate": 0.00016096997690531178,
-      "loss": 1.3155,
+      "loss": 0.0,
       "step": 1025
     },
     {
       "epoch": 0.4,
       "learning_rate": 0.00016077752117013086,
-      "loss": 1.2864,
+      "loss": 0.0,
       "step": 1030
     },
     {
       "epoch": 0.4,
       "learning_rate": 0.00016058506543494997,
-      "loss": 1.364,
+      "loss": 0.0,
       "step": 1035
     },
     {
       "epoch": 0.4,
       "learning_rate": 0.00016039260969976906,
-      "loss": 1.2698,
+      "loss": 0.0,
       "step": 1040
     },
     {
       "epoch": 0.4,
       "learning_rate": 0.00016020015396458817,
-      "loss": 1.3632,
+      "loss": 0.0,
       "step": 1045
     },
     {
       "epoch": 0.4,
       "learning_rate": 0.00016000769822940725,
-      "loss": 1.2741,
+      "loss": 0.0,
       "step": 1050
     },
     {
       "epoch": 0.41,
       "learning_rate": 0.00015981524249422633,
-      "loss": 1.3533,
+      "loss": 0.0,
       "step": 1055
     },
     {
       "epoch": 0.41,
       "learning_rate": 0.00015962278675904542,
-      "loss": 1.3571,
+      "loss": 0.0,
       "step": 1060
     },
     {
       "epoch": 0.41,
       "learning_rate": 0.0001594303310238645,
-      "loss": 1.2827,
+      "loss": 0.0,
       "step": 1065
     },
     {
       "epoch": 0.41,
       "learning_rate": 0.0001592378752886836,
-      "loss": 1.2547,
+      "loss": 0.0,
       "step": 1070
     },
     {
       "epoch": 0.41,
       "learning_rate": 0.0001590454195535027,
-      "loss": 1.4101,
+      "loss": 0.0,
       "step": 1075
     },
     {
       "epoch": 0.41,
       "learning_rate": 0.0001588529638183218,
-      "loss": 1.3149,
+      "loss": 0.0,
       "step": 1080
     },
     {
       "epoch": 0.42,
       "learning_rate": 0.00015866050808314088,
-      "loss": 1.3008,
+      "loss": 0.0,
       "step": 1085
     },
     {
       "epoch": 0.42,
       "learning_rate": 0.00015846805234796,
-      "loss": 1.2859,
+      "loss": 0.0,
       "step": 1090
     },
     {
       "epoch": 0.42,
       "learning_rate": 0.00015827559661277908,
-      "loss": 1.1892,
+      "loss": 0.0,
       "step": 1095
     },
     {
       "epoch": 0.42,
       "learning_rate": 0.00015808314087759816,
-      "loss": 1.364,
+      "loss": 0.0,
       "step": 1100
     },
     {
       "epoch": 0.42,
       "learning_rate": 0.00015789068514241724,
-      "loss": 1.2708,
+      "loss": 0.0,
       "step": 1105
     },
     {
       "epoch": 0.43,
       "learning_rate": 0.00015769822940723635,
-      "loss": 1.3591,
+      "loss": 0.0,
       "step": 1110
     },
     {
       "epoch": 0.43,
       "learning_rate": 0.00015750577367205543,
-      "loss": 1.2828,
+      "loss": 0.0,
       "step": 1115
     },
     {
       "epoch": 0.43,
       "learning_rate": 0.00015731331793687452,
-      "loss": 1.3861,
+      "loss": 0.0,
       "step": 1120
     },
     {
       "epoch": 0.43,
       "learning_rate": 0.00015712086220169363,
-      "loss": 1.3752,
+      "loss": 0.0,
       "step": 1125
     },
     {
       "epoch": 0.43,
       "learning_rate": 0.0001569284064665127,
-      "loss": 1.299,
+      "loss": 0.0,
       "step": 1130
     },
     {
       "epoch": 0.44,
       "learning_rate": 0.0001567359507313318,
-      "loss": 1.2744,
+      "loss": 0.0,
       "step": 1135
     },
     {
       "epoch": 0.44,
       "learning_rate": 0.00015654349499615087,
-      "loss": 1.3226,
+      "loss": 0.0,
       "step": 1140
     },
     {
       "epoch": 0.44,
       "learning_rate": 0.00015635103926096998,
-      "loss": 1.2342,
+      "loss": 0.0,
       "step": 1145
     },
     {
       "epoch": 0.44,
       "learning_rate": 0.00015615858352578907,
-      "loss": 1.2236,
+      "loss": 0.0,
       "step": 1150
     },
     {
       "epoch": 0.44,
       "learning_rate": 0.00015596612779060818,
-      "loss": 1.3178,
+      "loss": 0.0,
       "step": 1155
     },
     {
       "epoch": 0.45,
       "learning_rate": 0.00015577367205542726,
-      "loss": 1.315,
+      "loss": 0.0,
       "step": 1160
     },
     {
       "epoch": 0.45,
       "learning_rate": 0.00015558121632024637,
-      "loss": 1.2071,
+      "loss": 0.0,
       "step": 1165
     },
     {
       "epoch": 0.45,
       "learning_rate": 0.00015538876058506545,
-      "loss": 1.2485,
+      "loss": 0.0,
       "step": 1170
     },
     {
       "epoch": 0.45,
       "learning_rate": 0.00015519630484988454,
-      "loss": 1.3034,
+      "loss": 0.0,
       "step": 1175
     },
     {
       "epoch": 0.45,
       "learning_rate": 0.00015500384911470362,
-      "loss": 1.3113,
+      "loss": 0.0,
       "step": 1180
     },
     {
       "epoch": 0.46,
       "learning_rate": 0.0001548113933795227,
-      "loss": 1.4071,
+      "loss": 0.0,
       "step": 1185
     },
     {
       "epoch": 0.46,
       "learning_rate": 0.0001546189376443418,
-      "loss": 1.2806,
+      "loss": 0.0,
       "step": 1190
     },
     {
       "epoch": 0.46,
       "learning_rate": 0.0001544264819091609,
-      "loss": 1.2668,
+      "loss": 0.0,
       "step": 1195
     },
     {
       "epoch": 0.46,
       "learning_rate": 0.00015423402617398,
-      "loss": 1.2156,
+      "loss": 0.0,
       "step": 1200
     },
     {
       "epoch": 0.46,
       "learning_rate": 0.00015404157043879909,
-      "loss": 1.2713,
+      "loss": 0.0,
       "step": 1205
     },
     {
       "epoch": 0.46,
       "learning_rate": 0.00015384911470361817,
-      "loss": 1.3547,
+      "loss": 0.0,
       "step": 1210
     },
     {
       "epoch": 0.47,
       "learning_rate": 0.00015365665896843725,
-      "loss": 1.352,
+      "loss": 0.0,
       "step": 1215
     },
     {
       "epoch": 0.47,
       "learning_rate": 0.00015346420323325636,
-      "loss": 1.2166,
+      "loss": 0.0,
       "step": 1220
     },
     {
       "epoch": 0.47,
       "learning_rate": 0.00015327174749807544,
-      "loss": 1.3411,
+      "loss": 0.0,
       "step": 1225
     },
     {
       "epoch": 0.47,
       "learning_rate": 0.00015307929176289455,
-      "loss": 1.3127,
+      "loss": 0.0,
       "step": 1230
     },
     {
       "epoch": 0.47,
       "learning_rate": 0.00015288683602771364,
-      "loss": 1.2179,
+      "loss": 0.0,
       "step": 1235
     },
     {
       "epoch": 0.48,
       "learning_rate": 0.00015269438029253275,
-      "loss": 1.3099,
+      "loss": 0.0,
       "step": 1240
     },
     {
       "epoch": 0.48,
       "learning_rate": 0.00015250192455735183,
-      "loss": 1.3071,
+      "loss": 0.0,
       "step": 1245
     },
     {
       "epoch": 0.48,
       "learning_rate": 0.00015230946882217089,
-      "loss": 1.2789,
+      "loss": 0.0,
       "step": 1250
     },
     {
       "epoch": 0.48,
       "learning_rate": 0.00015211701308699,
-      "loss": 1.265,
+      "loss": 0.0,
       "step": 1255
     },
     {
       "epoch": 0.48,
       "learning_rate": 0.00015192455735180908,
-      "loss": 1.3423,
+      "loss": 0.0,
       "step": 1260
     },
     {
       "epoch": 0.49,
       "learning_rate": 0.0001517321016166282,
-      "loss": 1.3561,
+      "loss": 0.0,
       "step": 1265
     },
     {
       "epoch": 0.49,
       "learning_rate": 0.00015153964588144727,
-      "loss": 1.3766,
+      "loss": 0.0,
       "step": 1270
     },
     {
       "epoch": 0.49,
       "learning_rate": 0.00015134719014626638,
-      "loss": 1.1815,
+      "loss": 0.0,
       "step": 1275
     },
     {
       "epoch": 0.49,
       "learning_rate": 0.00015115473441108546,
-      "loss": 1.2731,
+      "loss": 0.0,
       "step": 1280
     },
     {
       "epoch": 0.49,
       "learning_rate": 0.00015096227867590455,
-      "loss": 1.4211,
+      "loss": 0.0,
       "step": 1285
     },
     {
       "epoch": 0.5,
       "learning_rate": 0.00015076982294072363,
-      "loss": 1.3037,
+      "loss": 0.0,
       "step": 1290
     },
     {
       "epoch": 0.5,
       "learning_rate": 0.00015057736720554274,
-      "loss": 1.2689,
+      "loss": 0.0,
       "step": 1295
     },
     {
       "epoch": 0.5,
       "learning_rate": 0.00015038491147036182,
-      "loss": 1.3821,
+      "loss": 0.0,
       "step": 1300
     },
     {
       "epoch": 0.5,
       "learning_rate": 0.00015019245573518093,
-      "loss": 1.3421,
+      "loss": 0.0,
       "step": 1305
     },
     {
       "epoch": 0.5,
       "learning_rate": 0.00015000000000000001,
-      "loss": 1.2462,
+      "loss": 0.0,
       "step": 1310
     },
     {
       "epoch": 0.51,
       "learning_rate": 0.0001498075442648191,
-      "loss": 1.3296,
+      "loss": 0.0,
       "step": 1315
     },
     {
       "epoch": 0.51,
       "learning_rate": 0.00014961508852963818,
-      "loss": 1.2991,
+      "loss": 0.0,
       "step": 1320
     },
     {
       "epoch": 0.51,
       "learning_rate": 0.00014942263279445726,
-      "loss": 1.2695,
+      "loss": 0.0,
       "step": 1325
     },
     {
       "epoch": 0.51,
       "learning_rate": 0.00014923017705927637,
-      "loss": 1.3681,
+      "loss": 0.0,
       "step": 1330
     },
     {
       "epoch": 0.51,
       "learning_rate": 0.00014903772132409545,
-      "loss": 1.3166,
+      "loss": 0.0,
       "step": 1335
     },
     {
       "epoch": 0.51,
       "learning_rate": 0.00014884526558891456,
-      "loss": 1.2906,
+      "loss": 0.0,
       "step": 1340
     },
     {
       "epoch": 0.52,
       "learning_rate": 0.00014865280985373365,
-      "loss": 1.2122,
+      "loss": 0.0,
       "step": 1345
     },
     {
       "epoch": 0.52,
       "learning_rate": 0.00014846035411855276,
-      "loss": 1.2767,
+      "loss": 0.0,
       "step": 1350
     },
     {
       "epoch": 0.52,
       "learning_rate": 0.00014826789838337184,
-      "loss": 1.3143,
+      "loss": 0.0,
       "step": 1355
     },
     {
       "epoch": 0.52,
       "learning_rate": 0.00014807544264819092,
-      "loss": 1.2427,
+      "loss": 0.0,
       "step": 1360
     },
     {
       "epoch": 0.52,
       "learning_rate": 0.00014788298691301,
-      "loss": 1.2413,
+      "loss": 0.0,
       "step": 1365
     },
     {
       "epoch": 0.53,
       "learning_rate": 0.00014769053117782912,
-      "loss": 1.2943,
+      "loss": 0.0,
       "step": 1370
     },
     {
       "epoch": 0.53,
       "learning_rate": 0.0001474980754426482,
-      "loss": 1.2527,
+      "loss": 0.0,
       "step": 1375
     },
     {
       "epoch": 0.53,
       "learning_rate": 0.00014730561970746728,
-      "loss": 1.3178,
+      "loss": 0.0,
       "step": 1380
     },
     {
       "epoch": 0.53,
       "learning_rate": 0.0001471131639722864,
-      "loss": 1.2924,
+      "loss": 0.0,
       "step": 1385
     },
     {
       "epoch": 0.53,
       "learning_rate": 0.00014692070823710547,
-      "loss": 1.1681,
+      "loss": 0.0,
       "step": 1390
     },
     {
       "epoch": 0.54,
       "learning_rate": 0.00014672825250192456,
-      "loss": 1.2805,
+      "loss": 0.0,
       "step": 1395
     },
     {
       "epoch": 0.54,
       "learning_rate": 0.00014653579676674364,
-      "loss": 1.2903,
+      "loss": 0.0,
       "step": 1400
     },
     {
       "epoch": 0.54,
       "learning_rate": 0.00014634334103156275,
-      "loss": 1.3223,
+      "loss": 0.0,
       "step": 1405
     },
     {
       "epoch": 0.54,
       "learning_rate": 0.00014615088529638183,
-      "loss": 1.208,
+      "loss": 0.0,
       "step": 1410
     },
     {
       "epoch": 0.54,
       "learning_rate": 0.00014595842956120094,
-      "loss": 1.2135,
+      "loss": 0.0,
       "step": 1415
     },
     {
       "epoch": 0.55,
       "learning_rate": 0.00014576597382602002,
-      "loss": 1.291,
+      "loss": 0.0,
       "step": 1420
     },
     {
       "epoch": 0.55,
       "learning_rate": 0.00014557351809083913,
-      "loss": 1.1809,
+      "loss": 0.0,
       "step": 1425
     },
     {
       "epoch": 0.55,
       "learning_rate": 0.00014538106235565822,
-      "loss": 1.3176,
+      "loss": 0.0,
       "step": 1430
     },
     {
       "epoch": 0.55,
       "learning_rate": 0.0001451886066204773,
-      "loss": 1.2837,
+      "loss": 0.0,
       "step": 1435
     },
     {
       "epoch": 0.55,
       "learning_rate": 0.00014499615088529638,
-      "loss": 1.2517,
+      "loss": 0.0,
       "step": 1440
     },
     {
       "epoch": 0.56,
       "learning_rate": 0.00014480369515011547,
-      "loss": 1.3029,
+      "loss": 0.0,
       "step": 1445
     },
     {
       "epoch": 0.56,
       "learning_rate": 0.00014461123941493458,
-      "loss": 1.2432,
+      "loss": 0.0,
       "step": 1450
     },
     {
       "epoch": 0.56,
       "learning_rate": 0.00014441878367975366,
-      "loss": 1.2917,
+      "loss": 0.0,
       "step": 1455
     },
     {
       "epoch": 0.56,
       "learning_rate": 0.00014422632794457277,
-      "loss": 1.3542,
+      "loss": 0.0,
       "step": 1460
     },
     {
       "epoch": 0.56,
       "learning_rate": 0.00014403387220939185,
-      "loss": 1.2598,
+      "loss": 0.0,
       "step": 1465
     },
     {
       "epoch": 0.56,
       "learning_rate": 0.00014384141647421093,
-      "loss": 1.2475,
+      "loss": 0.0,
       "step": 1470
     },
     {
       "epoch": 0.57,
       "learning_rate": 0.00014364896073903002,
-      "loss": 1.2112,
+      "loss": 0.0,
       "step": 1475
     },
     {
       "epoch": 0.57,
       "learning_rate": 0.00014345650500384913,
-      "loss": 1.2839,
+      "loss": 0.0,
       "step": 1480
     },
     {
       "epoch": 0.57,
       "learning_rate": 0.0001432640492686682,
-      "loss": 1.1871,
+      "loss": 0.0,
       "step": 1485
     },
     {
       "epoch": 0.57,
       "learning_rate": 0.00014307159353348732,
-      "loss": 1.1936,
+      "loss": 0.0,
       "step": 1490
     },
     {
       "epoch": 0.57,
       "learning_rate": 0.0001428791377983064,
-      "loss": 1.2619,
+      "loss": 0.0,
       "step": 1495
     },
     {
       "epoch": 0.58,
       "learning_rate": 0.0001426866820631255,
-      "loss": 1.243,
+      "loss": 0.0,
       "step": 1500
     },
     {
       "epoch": 0.58,
       "learning_rate": 0.00014249422632794457,
-      "loss": 1.3483,
+      "loss": 0.0,
       "step": 1505
     },
     {
       "epoch": 0.58,
       "learning_rate": 0.00014230177059276365,
-      "loss": 1.241,
+      "loss": 0.0,
       "step": 1510
     },
     {
       "epoch": 0.58,
       "learning_rate": 0.00014210931485758276,
-      "loss": 1.2585,
+      "loss": 0.0,
       "step": 1515
     },
     {
       "epoch": 0.58,
       "learning_rate": 0.00014191685912240184,
-      "loss": 1.2525,
+      "loss": 0.0,
       "step": 1520
     },
     {
       "epoch": 0.59,
       "learning_rate": 0.00014172440338722095,
-      "loss": 1.2934,
+      "loss": 0.0,
       "step": 1525
     },
     {
       "epoch": 0.59,
       "learning_rate": 0.00014153194765204003,
-      "loss": 1.2277,
+      "loss": 0.0,
       "step": 1530
     },
     {
       "epoch": 0.59,
       "learning_rate": 0.00014133949191685914,
-      "loss": 1.2294,
+      "loss": 0.0,
       "step": 1535
     },
     {
       "epoch": 0.59,
       "learning_rate": 0.00014114703618167823,
-      "loss": 1.245,
+      "loss": 0.0,
       "step": 1540
     },
     {
       "epoch": 0.59,
       "learning_rate": 0.0001409545804464973,
-      "loss": 1.1826,
+      "loss": 0.0,
       "step": 1545
     },
     {
       "epoch": 0.6,
       "learning_rate": 0.0001407621247113164,
-      "loss": 1.2436,
+      "loss": 0.0,
       "step": 1550
     },
     {
       "epoch": 0.6,
       "learning_rate": 0.0001405696689761355,
-      "loss": 1.1588,
+      "loss": 0.0,
       "step": 1555
     },
     {
       "epoch": 0.6,
       "learning_rate": 0.00014037721324095459,
-      "loss": 1.3642,
+      "loss": 0.0,
       "step": 1560
     },
     {
       "epoch": 0.6,
       "learning_rate": 0.0001401847575057737,
-      "loss": 1.2621,
+      "loss": 0.0,
       "step": 1565
     },
     {
       "epoch": 0.6,
       "learning_rate": 0.00013999230177059278,
-      "loss": 1.2909,
+      "loss": 0.0,
       "step": 1570
     },
     {
       "epoch": 0.61,
       "learning_rate": 0.00013979984603541186,
-      "loss": 1.2259,
+      "loss": 0.0,
       "step": 1575
     },
     {
       "epoch": 0.61,
       "learning_rate": 0.00013960739030023094,
-      "loss": 1.2078,
+      "loss": 0.0,
       "step": 1580
     },
     {
       "epoch": 0.61,
       "learning_rate": 0.00013941493456505003,
-      "loss": 1.2307,
+      "loss": 0.0,
       "step": 1585
     },
     {
       "epoch": 0.61,
       "learning_rate": 0.00013922247882986914,
-      "loss": 1.2816,
+      "loss": 0.0,
       "step": 1590
     },
     {
       "epoch": 0.61,
       "learning_rate": 0.00013903002309468822,
-      "loss": 1.2587,
+      "loss": 0.0,
       "step": 1595
     },
     {
       "epoch": 0.61,
       "learning_rate": 0.00013883756735950733,
-      "loss": 1.3018,
+      "loss": 0.0,
       "step": 1600
     },
     {
       "epoch": 0.62,
       "learning_rate": 0.0001386451116243264,
-      "loss": 1.2167,
+      "loss": 0.0,
       "step": 1605
     },
     {
       "epoch": 0.62,
       "learning_rate": 0.00013845265588914552,
-      "loss": 1.2399,
+      "loss": 0.0,
       "step": 1610
     },
     {
       "epoch": 0.62,
       "learning_rate": 0.0001382602001539646,
-      "loss": 1.2355,
+      "loss": 0.0,
       "step": 1615
     },
     {
       "epoch": 0.62,
       "learning_rate": 0.0001380677444187837,
-      "loss": 1.2099,
+      "loss": 0.0,
       "step": 1620
     },
     {
       "epoch": 0.62,
       "learning_rate": 0.00013787528868360277,
-      "loss": 1.2794,
+      "loss": 0.0,
       "step": 1625
     },
     {
       "epoch": 0.63,
       "learning_rate": 0.00013768283294842188,
-      "loss": 1.1232,
+      "loss": 0.0,
       "step": 1630
     },
     {
       "epoch": 0.63,
       "learning_rate": 0.00013749037721324096,
-      "loss": 1.2563,
+      "loss": 0.0,
       "step": 1635
     },
     {
       "epoch": 0.63,
       "learning_rate": 0.00013729792147806005,
-      "loss": 1.2364,
+      "loss": 0.0,
       "step": 1640
     },
     {
       "epoch": 0.63,
       "learning_rate": 0.00013710546574287916,
-      "loss": 1.2036,
+      "loss": 0.0,
       "step": 1645
     },
     {
       "epoch": 0.63,
       "learning_rate": 0.00013691301000769824,
-      "loss": 1.209,
+      "loss": 0.0,
       "step": 1650
     },
     {
       "epoch": 0.64,
       "learning_rate": 0.00013672055427251732,
-      "loss": 1.2653,
+      "loss": 0.0,
       "step": 1655
     },
     {
       "epoch": 0.64,
       "learning_rate": 0.0001365280985373364,
-      "loss": 1.3329,
+      "loss": 0.0,
       "step": 1660
     },
     {
       "epoch": 0.64,
       "learning_rate": 0.0001363356428021555,
-      "loss": 1.2629,
+      "loss": 0.0,
       "step": 1665
     },
     {
       "epoch": 0.64,
       "learning_rate": 0.0001361431870669746,
-      "loss": 1.2393,
+      "loss": 0.0,
       "step": 1670
     },
     {
       "epoch": 0.64,
       "learning_rate": 0.0001359507313317937,
-      "loss": 1.1669,
+      "loss": 0.0,
       "step": 1675
     },
     {
       "epoch": 0.65,
       "learning_rate": 0.0001357582755966128,
-      "loss": 1.2177,
+      "loss": 0.0,
       "step": 1680
     },
     {
       "epoch": 0.65,
       "learning_rate": 0.0001355658198614319,
-      "loss": 1.1363,
+      "loss": 0.0,
       "step": 1685
     },
     {
       "epoch": 0.65,
       "learning_rate": 0.00013537336412625098,
-      "loss": 1.1773,
+      "loss": 0.0,
       "step": 1690
     },
     {
       "epoch": 0.65,
       "learning_rate": 0.00013518090839107006,
-      "loss": 1.2482,
+      "loss": 0.0,
       "step": 1695
     },
     {
       "epoch": 0.65,
       "learning_rate": 0.00013498845265588915,
-      "loss": 1.266,
+      "loss": 0.0,
       "step": 1700
     },
     {
       "epoch": 0.66,
       "learning_rate": 0.00013479599692070823,
-      "loss": 1.2936,
+      "loss": 0.0,
       "step": 1705
     },
     {
       "epoch": 0.66,
       "learning_rate": 0.00013460354118552734,
-      "loss": 1.2855,
+      "loss": 0.0,
       "step": 1710
     },
     {
       "epoch": 0.66,
       "learning_rate": 0.00013441108545034642,
-      "loss": 1.2413,
+      "loss": 0.0,
       "step": 1715
     },
     {
       "epoch": 0.66,
       "learning_rate": 0.00013421862971516553,
-      "loss": 1.2874,
+      "loss": 0.0,
       "step": 1720
     },
     {
       "epoch": 0.66,
       "learning_rate": 0.00013402617397998461,
-      "loss": 1.2527,
+      "loss": 0.0,
       "step": 1725
     },
     {
       "epoch": 0.66,
       "learning_rate": 0.0001338337182448037,
-      "loss": 1.2527,
+      "loss": 0.0,
       "step": 1730
     },
     {
       "epoch": 0.67,
       "learning_rate": 0.00013364126250962278,
-      "loss": 1.2423,
+      "loss": 0.0,
       "step": 1735
     },
     {
       "epoch": 0.67,
       "learning_rate": 0.0001334488067744419,
-      "loss": 1.2544,
+      "loss": 0.0,
       "step": 1740
     },
     {
       "epoch": 0.67,
       "learning_rate": 0.00013325635103926097,
-      "loss": 1.1613,
+      "loss": 0.0,
       "step": 1745
     },
     {
       "epoch": 0.67,
       "learning_rate": 0.00013306389530408008,
-      "loss": 1.2273,
+      "loss": 0.0,
       "step": 1750
     },
     {
       "epoch": 0.67,
       "learning_rate": 0.00013287143956889917,
-      "loss": 1.2789,
+      "loss": 0.0,
       "step": 1755
     },
     {
       "epoch": 0.68,
       "learning_rate": 0.00013267898383371828,
-      "loss": 1.2755,
+      "loss": 0.0,
       "step": 1760
     },
     {
       "epoch": 0.68,
       "learning_rate": 0.00013248652809853733,
-      "loss": 1.2144,
+      "loss": 0.0,
       "step": 1765
     },
     {
       "epoch": 0.68,
       "learning_rate": 0.00013229407236335641,
-      "loss": 1.1615,
+      "loss": 0.0,
       "step": 1770
     },
     {
       "epoch": 0.68,
       "learning_rate": 0.00013210161662817552,
-      "loss": 1.2173,
+      "loss": 0.0,
       "step": 1775
     },
     {
       "epoch": 0.68,
       "learning_rate": 0.0001319091608929946,
-      "loss": 1.2408,
+      "loss": 0.0,
       "step": 1780
     },
     {
       "epoch": 0.69,
       "learning_rate": 0.00013171670515781372,
-      "loss": 1.1671,
+      "loss": 0.0,
       "step": 1785
     },
     {
       "epoch": 0.69,
       "learning_rate": 0.0001315242494226328,
-      "loss": 1.1994,
+      "loss": 0.0,
       "step": 1790
     },
     {
       "epoch": 0.69,
       "learning_rate": 0.0001313317936874519,
-      "loss": 1.3369,
+      "loss": 0.0,
       "step": 1795
     },
     {
       "epoch": 0.69,
       "learning_rate": 0.000131139337952271,
-      "loss": 1.1598,
+      "loss": 0.0,
       "step": 1800
     },
     {
       "epoch": 0.69,
       "learning_rate": 0.00013094688221709007,
-      "loss": 1.1784,
+      "loss": 0.0,
       "step": 1805
     },
     {
       "epoch": 0.7,
       "learning_rate": 0.00013075442648190916,
-      "loss": 1.2743,
+      "loss": 0.0,
       "step": 1810
     },
     {
       "epoch": 0.7,
       "learning_rate": 0.00013056197074672827,
-      "loss": 1.2127,
+      "loss": 0.0,
       "step": 1815
     },
     {
       "epoch": 0.7,
       "learning_rate": 0.00013036951501154735,
-      "loss": 1.2738,
+      "loss": 0.0,
       "step": 1820
     },
     {
       "epoch": 0.7,
       "learning_rate": 0.00013017705927636643,
-      "loss": 1.2232,
+      "loss": 0.0,
       "step": 1825
     },
     {
       "epoch": 0.7,
       "learning_rate": 0.00012998460354118554,
-      "loss": 1.1547,
+      "loss": 0.0,
       "step": 1830
     },
     {
       "epoch": 0.7,
       "learning_rate": 0.00012979214780600463,
-      "loss": 1.2235,
+      "loss": 0.0,
       "step": 1835
     },
     {
       "epoch": 0.71,
       "learning_rate": 0.0001295996920708237,
-      "loss": 1.332,
+      "loss": 0.0,
       "step": 1840
     },
     {
       "epoch": 0.71,
       "learning_rate": 0.0001294072363356428,
-      "loss": 1.1242,
+      "loss": 0.0,
       "step": 1845
     },
     {
       "epoch": 0.71,
       "learning_rate": 0.0001292147806004619,
-      "loss": 1.1874,
+      "loss": 0.0,
       "step": 1850
     },
     {
       "epoch": 0.71,
       "learning_rate": 0.00012902232486528098,
-      "loss": 1.1394,
+      "loss": 0.0,
       "step": 1855
     },
     {
       "epoch": 0.71,
       "learning_rate": 0.0001288298691301001,
-      "loss": 1.1734,
+      "loss": 0.0,
       "step": 1860
     },
     {
       "epoch": 0.72,
       "learning_rate": 0.00012863741339491918,
-      "loss": 1.2245,
+      "loss": 0.0,
       "step": 1865
     },
     {
       "epoch": 0.72,
       "learning_rate": 0.00012844495765973829,
-      "loss": 1.1869,
+      "loss": 0.0,
       "step": 1870
     },
     {
       "epoch": 0.72,
       "learning_rate": 0.00012825250192455737,
-      "loss": 1.2558,
+      "loss": 0.0,
       "step": 1875
     }
   ],
   "max_steps": 5206,
   "num_train_epochs": 2,
-  "total_flos": 19105201520640.0,
+  "total_flos": 28526020919296.0,
   "trial_name": null,
   "trial_params": null
 }

last-checkpoint/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:762f87ade0fa6168b36ab9775461c0f8a3950e947fee8d4e9ec9e059b0407a8a
+oid sha256:06fa3549f7811417470fddb9a68287828d541a4c77fedca8c058a99d9790f5d0
 size 6139

special_tokens_map.json

@@ -1,26 +1,6 @@
 {
-  "bos_token": {
-    "content": "<s>",
-    "lstrip": false,
-    "normalized": true,
-    "rstrip": false,
-    "single_word": false
-  },
-  "cls_token": "</s>",
-  "eos_token": {
-    "content": "</s>",
-    "lstrip": false,
-    "normalized": true,
-    "rstrip": false,
-    "single_word": false
-  },
+  "bos_token": "<s>",
+  "eos_token": "</s>",
   "pad_token": "</s>",
-  "sep_token": "</s>",
-  "unk_token": {
-    "content": "<unk>",
-    "lstrip": false,
-    "normalized": true,
-    "rstrip": false,
-    "single_word": false
-  }
+  "unk_token": "<unk>"
 }

tokenizer.model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ab1b681ec7fc02fed5edd3026687d7a692a918c4dd8e150ca2e3994a6229843b
-size 534194
+oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
+size 499723

tokenizer_config.json

@@ -1,34 +1,14 @@
 {
-  "add_bos_token": true,
-  "add_eos_token": false,
-  "add_prefix_space": false,
-  "bos_token": {
-    "__type": "AddedToken",
-    "content": "<s>",
-    "lstrip": false,
-    "normalized": true,
-    "rstrip": false,
-    "single_word": false
-  },
+  "bos_token": "<s>",
   "clean_up_tokenization_spaces": false,
-  "eos_token": {
-    "__type": "AddedToken",
-    "content": "</s>",
-    "lstrip": false,
-    "normalized": true,
-    "rstrip": false,
-    "single_word": false
-  },
-  "model_max_length": 2048,
+  "device_map": "cuda",
+  "eos_token": "</s>",
+  "max_length": 256,
+  "model_max_length": 1000000000000000019884624838656,
   "pad_token": null,
+  "padding": "max_length",
   "sp_model_kwargs": {},
   "tokenizer_class": "LlamaTokenizer",
-  "unk_token": {
-    "__type": "AddedToken",
-    "content": "<unk>",
-    "lstrip": false,
-    "normalized": true,
-    "rstrip": false,
-    "single_word": false
-  }
+  "truncation": "longest_first",
+  "unk_token": "<unk>"
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:762f87ade0fa6168b36ab9775461c0f8a3950e947fee8d4e9ec9e059b0407a8a
+oid sha256:06fa3549f7811417470fddb9a68287828d541a4c77fedca8c058a99d9790f5d0
 size 6139