Upload folder using huggingface_hub

checkpoint-step-1200/adapter_config.json

@@ -34,13 +34,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
+    "gate_proj",
+    "k_proj",
     "v_proj",
     "down_proj",
     "o_proj",
-    "q_proj",
     "up_proj",
-    "gate_proj",
-    "k_proj"
+    "q_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",

checkpoint-step-1200/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6a0638027abedc4c2269d44ea45c717c23688c734781360cd8f356e2b58856cf
+oid sha256:53364c92578b73ee024ed722471a475ab20e95a9d865269c30543abcfec994f8
 size 239536272

checkpoint-step-1200/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0607d696320cbb9c966684bce0a7e12f4cd6504d1754719140d4088458e0d6ed
+oid sha256:8fd70155a7042c6f2348e03114ba317896403d65ad528fb3b5e6d2e45a9674f3
 size 122203333

checkpoint-step-1200/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5ec619f50d6dc085f5fa346c2e7bad8f96ad5b3bb7557c76b686b76aaf1c3509
+oid sha256:b9466d2afca7fd6124ef1c1eb08fe1d29700db0d98947cd5aaae851b7a15434c
 size 1465

checkpoint-step-1200/tokenizer_config.json

@@ -5,7 +5,7 @@
   "clean_up_tokenization_spaces": false,
   "eos_token": "<|endoftext|>",
   "errors": "replace",
-  "is_local": false,
+  "is_local": true,
   "model_max_length": 32768,
   "pad_token": "<|PAD_TOKEN|>",
   "padding_side": "right",

checkpoint-step-1200/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_global_step": 1000,
-  "best_metric": 0.9810675382614136,
-  "best_model_checkpoint": "./outputs/checkpoint-1000",
+  "best_global_step": 400,
+  "best_metric": 0.9676186442375183,
+  "best_model_checkpoint": "./outputs/checkpoint-400",
   "epoch": 2.255639097744361,
   "eval_steps": 100,
   "global_step": 1200,
@@ -11,273 +11,273 @@
   "log_history": [
     {
       "epoch": 0.09398496240601503,
-      "grad_norm": 0.49556756019592285,
+      "grad_norm": 0.9101622104644775,
       "learning_rate": 9.8e-05,
-      "loss": 1.8540109252929688,
+      "loss": 0.7848518371582032,
       "step": 50
     },
     {
       "epoch": 0.18796992481203006,
-      "grad_norm": 0.4922962486743927,
-      "learning_rate": 9.907020872865275e-05,
-      "loss": 1.2944253540039063,
+      "grad_norm": 0.8308438062667847,
+      "learning_rate": 9.683053040103493e-05,
+      "loss": 0.7757094573974609,
       "step": 100
     },
     {
       "epoch": 0.18796992481203006,
-      "eval_loss": 1.216850757598877,
-      "eval_runtime": 83.4733,
-      "eval_samples_per_second": 5.666,
-      "eval_steps_per_second": 1.426,
+      "eval_loss": 0.9840108752250671,
+      "eval_runtime": 81.3507,
+      "eval_samples_per_second": 5.814,
+      "eval_steps_per_second": 1.463,
       "step": 100
     },
     {
       "epoch": 0.2819548872180451,
-      "grad_norm": 0.5891006588935852,
-      "learning_rate": 9.81214421252372e-05,
-      "loss": 1.2092308807373047,
+      "grad_norm": 1.0474584102630615,
+      "learning_rate": 9.359637774902976e-05,
+      "loss": 0.7674188995361328,
       "step": 150
     },
     {
       "epoch": 0.37593984962406013,
-      "grad_norm": 0.5563158392906189,
-      "learning_rate": 9.717267552182164e-05,
-      "loss": 1.1433011627197265,
+      "grad_norm": 0.8780233860015869,
+      "learning_rate": 9.036222509702458e-05,
+      "loss": 0.7395731353759766,
       "step": 200
     },
     {
       "epoch": 0.37593984962406013,
-      "eval_loss": 1.1367270946502686,
-      "eval_runtime": 82.9022,
-      "eval_samples_per_second": 5.706,
-      "eval_steps_per_second": 1.435,
+      "eval_loss": 0.9940316081047058,
+      "eval_runtime": 80.2928,
+      "eval_samples_per_second": 5.891,
+      "eval_steps_per_second": 1.482,
       "step": 200
     },
     {
       "epoch": 0.4699248120300752,
-      "grad_norm": 0.5556582808494568,
-      "learning_rate": 9.622390891840608e-05,
-      "loss": 1.1522552490234375,
+      "grad_norm": 0.9176077246665955,
+      "learning_rate": 8.712807244501941e-05,
+      "loss": 0.7627955627441406,
       "step": 250
     },
     {
       "epoch": 0.5639097744360902,
-      "grad_norm": 0.5758287906646729,
-      "learning_rate": 9.527514231499052e-05,
-      "loss": 1.0814830780029296,
+      "grad_norm": 0.8625491857528687,
+      "learning_rate": 8.389391979301423e-05,
+      "loss": 0.7164928436279296,
       "step": 300
     },
     {
       "epoch": 0.5639097744360902,
-      "eval_loss": 1.1028286218643188,
-      "eval_runtime": 82.616,
-      "eval_samples_per_second": 5.725,
-      "eval_steps_per_second": 1.44,
+      "eval_loss": 0.9823606014251709,
+      "eval_runtime": 80.8571,
+      "eval_samples_per_second": 5.85,
+      "eval_steps_per_second": 1.472,
       "step": 300
     },
     {
       "epoch": 0.6578947368421053,
-      "grad_norm": 0.5769486427307129,
-      "learning_rate": 9.432637571157496e-05,
-      "loss": 1.124346466064453,
+      "grad_norm": 0.8384292125701904,
+      "learning_rate": 8.065976714100906e-05,
+      "loss": 0.73987548828125,
       "step": 350
     },
     {
       "epoch": 0.7518796992481203,
-      "grad_norm": 0.5422045588493347,
-      "learning_rate": 9.337760910815939e-05,
-      "loss": 1.0794380950927733,
+      "grad_norm": 0.8384448885917664,
+      "learning_rate": 7.742561448900388e-05,
+      "loss": 0.7277918243408203,
       "step": 400
     },
     {
       "epoch": 0.7518796992481203,
-      "eval_loss": 1.072836995124817,
-      "eval_runtime": 82.9563,
-      "eval_samples_per_second": 5.702,
-      "eval_steps_per_second": 1.434,
+      "eval_loss": 0.9676186442375183,
+      "eval_runtime": 80.4443,
+      "eval_samples_per_second": 5.88,
+      "eval_steps_per_second": 1.479,
       "step": 400
     },
     {
       "epoch": 0.8458646616541353,
-      "grad_norm": 0.5861876606941223,
-      "learning_rate": 9.242884250474384e-05,
-      "loss": 1.0750784301757812,
+      "grad_norm": 0.866162121295929,
+      "learning_rate": 7.41914618369987e-05,
+      "loss": 0.7323916625976562,
       "step": 450
     },
     {
       "epoch": 0.9398496240601504,
-      "grad_norm": 0.5859329700469971,
-      "learning_rate": 9.148007590132827e-05,
-      "loss": 1.083390655517578,
+      "grad_norm": 0.8226180076599121,
+      "learning_rate": 7.095730918499353e-05,
+      "loss": 0.7260687255859375,
       "step": 500
     },
     {
       "epoch": 0.9398496240601504,
-      "eval_loss": 1.0487685203552246,
-      "eval_runtime": 82.9841,
-      "eval_samples_per_second": 5.7,
-      "eval_steps_per_second": 1.434,
+      "eval_loss": 0.9678767919540405,
+      "eval_runtime": 80.5208,
+      "eval_samples_per_second": 5.874,
+      "eval_steps_per_second": 1.478,
       "step": 500
     },
     {
       "epoch": 1.0338345864661653,
-      "grad_norm": 0.787390410900116,
-      "learning_rate": 9.053130929791271e-05,
-      "loss": 1.0226116943359376,
+      "grad_norm": 1.0985685586929321,
+      "learning_rate": 6.772315653298837e-05,
+      "loss": 0.6786854553222657,
       "step": 550
     },
     {
       "epoch": 1.1278195488721805,
-      "grad_norm": 0.7412646412849426,
-      "learning_rate": 8.958254269449717e-05,
-      "loss": 0.9393210601806641,
+      "grad_norm": 1.0968552827835083,
+      "learning_rate": 6.44890038809832e-05,
+      "loss": 0.580701789855957,
       "step": 600
     },
     {
       "epoch": 1.1278195488721805,
-      "eval_loss": 1.0402475595474243,
-      "eval_runtime": 83.0546,
-      "eval_samples_per_second": 5.695,
-      "eval_steps_per_second": 1.433,
+      "eval_loss": 1.0256125926971436,
+      "eval_runtime": 80.6908,
+      "eval_samples_per_second": 5.862,
+      "eval_steps_per_second": 1.475,
       "step": 600
     },
     {
       "epoch": 1.2218045112781954,
-      "grad_norm": 0.6759052276611328,
-      "learning_rate": 8.86337760910816e-05,
-      "loss": 0.9417331695556641,
+      "grad_norm": 1.0653274059295654,
+      "learning_rate": 6.1254851228978e-05,
+      "loss": 0.583438491821289,
       "step": 650
     },
     {
       "epoch": 1.3157894736842106,
-      "grad_norm": 0.7094544172286987,
-      "learning_rate": 8.768500948766604e-05,
-      "loss": 0.9495692443847656,
+      "grad_norm": 1.1410211324691772,
+      "learning_rate": 5.802069857697283e-05,
+      "loss": 0.5911675262451171,
       "step": 700
     },
     {
       "epoch": 1.3157894736842106,
-      "eval_loss": 1.0223630666732788,
-      "eval_runtime": 82.8984,
-      "eval_samples_per_second": 5.706,
-      "eval_steps_per_second": 1.435,
+      "eval_loss": 1.0415176153182983,
+      "eval_runtime": 80.2742,
+      "eval_samples_per_second": 5.892,
+      "eval_steps_per_second": 1.482,
       "step": 700
     },
     {
       "epoch": 1.4097744360902256,
-      "grad_norm": 0.7613663077354431,
-      "learning_rate": 8.673624288425048e-05,
-      "loss": 0.9635891723632812,
+      "grad_norm": 1.0943447351455688,
+      "learning_rate": 5.478654592496766e-05,
+      "loss": 0.6138821029663086,
       "step": 750
     },
     {
       "epoch": 1.5037593984962405,
-      "grad_norm": 0.6828358173370361,
-      "learning_rate": 8.578747628083492e-05,
-      "loss": 0.9501354217529296,
+      "grad_norm": 1.004882574081421,
+      "learning_rate": 5.1552393272962485e-05,
+      "loss": 0.6031227111816406,
       "step": 800
     },
     {
       "epoch": 1.5037593984962405,
-      "eval_loss": 1.0105401277542114,
-      "eval_runtime": 82.7473,
-      "eval_samples_per_second": 5.716,
-      "eval_steps_per_second": 1.438,
+      "eval_loss": 1.0161856412887573,
+      "eval_runtime": 80.0684,
+      "eval_samples_per_second": 5.907,
+      "eval_steps_per_second": 1.486,
       "step": 800
     },
     {
       "epoch": 1.5977443609022557,
-      "grad_norm": 0.7703934907913208,
-      "learning_rate": 8.483870967741936e-05,
-      "loss": 0.9451417541503906,
+      "grad_norm": 1.0877087116241455,
+      "learning_rate": 4.831824062095731e-05,
+      "loss": 0.5991556930541992,
       "step": 850
     },
     {
       "epoch": 1.6917293233082706,
-      "grad_norm": 0.8639416098594666,
-      "learning_rate": 8.38899430740038e-05,
-      "loss": 0.9263732147216797,
+      "grad_norm": 1.2106397151947021,
+      "learning_rate": 4.508408796895214e-05,
+      "loss": 0.6013988876342773,
       "step": 900
     },
     {
       "epoch": 1.6917293233082706,
-      "eval_loss": 0.9950240254402161,
-      "eval_runtime": 82.6629,
-      "eval_samples_per_second": 5.722,
-      "eval_steps_per_second": 1.44,
+      "eval_loss": 1.0028849840164185,
+      "eval_runtime": 80.8781,
+      "eval_samples_per_second": 5.848,
+      "eval_steps_per_second": 1.471,
       "step": 900
     },
     {
       "epoch": 1.7857142857142856,
-      "grad_norm": 0.7719340324401855,
-      "learning_rate": 8.294117647058824e-05,
-      "loss": 0.9348648834228516,
+      "grad_norm": 1.123062252998352,
+      "learning_rate": 4.1849935316946965e-05,
+      "loss": 0.592793083190918,
       "step": 950
     },
     {
       "epoch": 1.8796992481203008,
-      "grad_norm": 0.6620801687240601,
-      "learning_rate": 8.199240986717268e-05,
-      "loss": 0.9322538757324219,
+      "grad_norm": 1.0219600200653076,
+      "learning_rate": 3.861578266494178e-05,
+      "loss": 0.6088079071044922,
       "step": 1000
     },
     {
       "epoch": 1.8796992481203008,
-      "eval_loss": 0.9810675382614136,
-      "eval_runtime": 82.874,
-      "eval_samples_per_second": 5.707,
-      "eval_steps_per_second": 1.436,
+      "eval_loss": 1.0209505558013916,
+      "eval_runtime": 80.909,
+      "eval_samples_per_second": 5.846,
+      "eval_steps_per_second": 1.471,
       "step": 1000
     },
     {
       "epoch": 1.973684210526316,
-      "grad_norm": 0.904947817325592,
-      "learning_rate": 8.104364326375712e-05,
-      "loss": 0.929292221069336,
+      "grad_norm": 1.03061044216156,
+      "learning_rate": 3.5381630012936615e-05,
+      "loss": 0.5962826919555664,
       "step": 1050
     },
     {
       "epoch": 2.0676691729323307,
-      "grad_norm": 0.7428781390190125,
-      "learning_rate": 8.009487666034155e-05,
-      "loss": 0.8347126770019532,
+      "grad_norm": 1.2498968839645386,
+      "learning_rate": 3.214747736093144e-05,
+      "loss": 0.47724185943603514,
       "step": 1100
     },
     {
       "epoch": 2.0676691729323307,
-      "eval_loss": 0.9958235621452332,
-      "eval_runtime": 82.8108,
-      "eval_samples_per_second": 5.712,
-      "eval_steps_per_second": 1.437,
+      "eval_loss": 1.1276904344558716,
+      "eval_runtime": 81.046,
+      "eval_samples_per_second": 5.836,
+      "eval_steps_per_second": 1.468,
       "step": 1100
     },
     {
       "epoch": 2.161654135338346,
-      "grad_norm": 0.8485831618309021,
-      "learning_rate": 7.9146110056926e-05,
-      "loss": 0.7888198852539062,
+      "grad_norm": 1.1484302282333374,
+      "learning_rate": 2.8913324708926264e-05,
+      "loss": 0.4027834701538086,
       "step": 1150
     },
     {
       "epoch": 2.255639097744361,
-      "grad_norm": 0.9114283323287964,
-      "learning_rate": 7.819734345351045e-05,
-      "loss": 0.7951007843017578,
+      "grad_norm": 1.2905447483062744,
+      "learning_rate": 2.5679172056921085e-05,
+      "loss": 0.4188160705566406,
       "step": 1200
     },
     {
       "epoch": 2.255639097744361,
-      "eval_loss": 0.9948464632034302,
-      "eval_runtime": 82.8049,
-      "eval_samples_per_second": 5.712,
-      "eval_steps_per_second": 1.437,
+      "eval_loss": 1.1084345579147339,
+      "eval_runtime": 80.5472,
+      "eval_samples_per_second": 5.872,
+      "eval_steps_per_second": 1.477,
       "step": 1200
     }
   ],
   "logging_steps": 50,
-  "max_steps": 5320,
+  "max_steps": 1596,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 10,
+  "num_train_epochs": 3,
   "save_steps": 100,
   "stateful_callbacks": {
     "TrainerControl": {

checkpoint-step-1200/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bc05806d2c86681ef6c541c75942126f28b3a31aec0d903899ed9bf1ea303bac
+oid sha256:3c88f5ec73ab38f2f0a5bb321f3c58aaa6f716ae0e803b69ff493f9e0c43f221
 size 5713