sync 2026-04-24T07:52:39+00:00

sft/checkpoint-400/adapter_config.json

@@ -30,8 +30,8 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "q_proj",
     "k_proj",
+    "q_proj",
     "v_proj",
     "o_proj"
   ],

sft/checkpoint-400/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b152714f3de9276f49eb38c48fa0911fb388519f0094cc4a1f17b7bbe1907129
+oid sha256:e4160a7f58c058e509163ce331774f03fdfa138d52a65bc19dcba71f10f0cf86
 size 61380432

sft/checkpoint-400/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:afd44d2d8ede5e8e25d59fe05c88b0f0f201b83fb64dd0ba80ed0f93c7cc52f1
+oid sha256:9d995674cea91aba25b767784162d650dc974d9d00e1673050234bd7b0db3ce2
 size 122930379

sft/checkpoint-400/trainer_state.json

@@ -11,282 +11,282 @@
   "log_history": [
     {
       "epoch": 0.04094165813715456,
-      "grad_norm": 6.233852386474609,
+      "grad_norm": 7.529000759124756,
       "learning_rate": 9.000000000000001e-07,
-      "loss": 6.206051254272461,
+      "loss": 6.639524841308594,
       "step": 10
     },
     {
       "epoch": 0.08188331627430911,
-      "grad_norm": 5.675178527832031,
+      "grad_norm": 7.1427388191223145,
       "learning_rate": 1.9000000000000002e-06,
-      "loss": 6.079831314086914,
+      "loss": 6.582770538330078,
       "step": 20
     },
     {
       "epoch": 0.12282497441146366,
-      "grad_norm": 5.132958889007568,
+      "grad_norm": 6.251121997833252,
       "learning_rate": 2.9e-06,
-      "loss": 5.749290466308594,
+      "loss": 6.114261245727539,
       "step": 30
     },
     {
       "epoch": 0.16376663254861823,
-      "grad_norm": 5.81948184967041,
+      "grad_norm": 7.014133453369141,
       "learning_rate": 3.900000000000001e-06,
-      "loss": 6.033080673217773,
+      "loss": 6.333962631225586,
       "step": 40
     },
     {
       "epoch": 0.2047082906857728,
-      "grad_norm": 6.442906379699707,
+      "grad_norm": 7.072969913482666,
       "learning_rate": 4.9000000000000005e-06,
-      "loss": 5.972381973266602,
+      "loss": 6.248543930053711,
       "step": 50
     },
     {
       "epoch": 0.24564994882292732,
-      "grad_norm": 5.304522514343262,
+      "grad_norm": 5.674424171447754,
       "learning_rate": 5.9e-06,
-      "loss": 5.472136306762695,
+      "loss": 5.6220745086669925,
       "step": 60
     },
     {
       "epoch": 0.2865916069600819,
-      "grad_norm": 5.261083126068115,
+      "grad_norm": 5.3739800453186035,
       "learning_rate": 6.9e-06,
-      "loss": 5.041688537597656,
+      "loss": 5.0251930236816404,
       "step": 70
     },
     {
       "epoch": 0.32753326509723646,
-      "grad_norm": 6.387202739715576,
+      "grad_norm": 6.372293472290039,
       "learning_rate": 7.9e-06,
-      "loss": 4.916830062866211,
+      "loss": 5.033017349243164,
       "step": 80
     },
     {
       "epoch": 0.368474923234391,
-      "grad_norm": 7.068334102630615,
+      "grad_norm": 7.22106409072876,
       "learning_rate": 8.900000000000001e-06,
-      "loss": 3.531498336791992,
+      "loss": 3.514456939697266,
       "step": 90
     },
     {
       "epoch": 0.4094165813715456,
-      "grad_norm": 4.787615776062012,
+      "grad_norm": 4.990958213806152,
       "learning_rate": 9.9e-06,
-      "loss": 2.2226877212524414,
+      "loss": 2.256962776184082,
       "step": 100
     },
     {
       "epoch": 0.4503582395087001,
-      "grad_norm": 3.4490268230438232,
+      "grad_norm": 3.2661993503570557,
       "learning_rate": 9.76923076923077e-06,
-      "loss": 1.955281639099121,
+      "loss": 2.0482555389404298,
       "step": 110
     },
     {
       "epoch": 0.49129989764585463,
-      "grad_norm": 2.435624361038208,
+      "grad_norm": 2.997323513031006,
       "learning_rate": 9.512820512820514e-06,
-      "loss": 1.5814408302307128,
+      "loss": 1.6077747344970703,
       "step": 120
     },
     {
       "epoch": 0.5322415557830092,
-      "grad_norm": 2.0870230197906494,
+      "grad_norm": 2.422741413116455,
       "learning_rate": 9.256410256410257e-06,
-      "loss": 1.4773900985717774,
+      "loss": 1.538028621673584,
       "step": 130
     },
     {
       "epoch": 0.5731832139201638,
-      "grad_norm": 2.1049163341522217,
+      "grad_norm": 2.4594290256500244,
       "learning_rate": 9e-06,
-      "loss": 1.508018970489502,
+      "loss": 1.5614049911499024,
       "step": 140
     },
     {
       "epoch": 0.6141248720573184,
-      "grad_norm": 2.7535016536712646,
+      "grad_norm": 2.7256577014923096,
       "learning_rate": 8.743589743589743e-06,
-      "loss": 1.6693695068359375,
+      "loss": 1.7180768966674804,
       "step": 150
     },
     {
       "epoch": 0.6550665301944729,
-      "grad_norm": 2.2710325717926025,
+      "grad_norm": 2.6614902019500732,
       "learning_rate": 8.487179487179488e-06,
-      "loss": 1.590962028503418,
+      "loss": 1.6412044525146485,
       "step": 160
     },
     {
       "epoch": 0.6960081883316275,
-      "grad_norm": 2.007903575897217,
+      "grad_norm": 2.1517934799194336,
       "learning_rate": 8.230769230769232e-06,
-      "loss": 1.6514751434326171,
+      "loss": 1.729467010498047,
       "step": 170
     },
     {
       "epoch": 0.736949846468782,
-      "grad_norm": 2.2933390140533447,
+      "grad_norm": 2.4588229656219482,
       "learning_rate": 7.974358974358975e-06,
-      "loss": 1.2684261322021484,
+      "loss": 1.305363941192627,
       "step": 180
     },
     {
       "epoch": 0.7778915046059366,
-      "grad_norm": 2.4877562522888184,
+      "grad_norm": 2.5282061100006104,
       "learning_rate": 7.717948717948718e-06,
-      "loss": 1.4245257377624512,
+      "loss": 1.4986873626708985,
       "step": 190
     },
     {
       "epoch": 0.8188331627430911,
-      "grad_norm": 2.4900834560394287,
+      "grad_norm": 3.699396848678589,
       "learning_rate": 7.461538461538462e-06,
-      "loss": 1.4417343139648438,
+      "loss": 1.5059691429138184,
       "step": 200
     },
     {
       "epoch": 0.8597748208802457,
-      "grad_norm": 1.8601328134536743,
+      "grad_norm": 2.2632241249084473,
       "learning_rate": 7.205128205128206e-06,
-      "loss": 1.3829060554504395,
+      "loss": 1.4249659538269044,
       "step": 210
     },
     {
       "epoch": 0.9007164790174002,
-      "grad_norm": 1.3816406726837158,
+      "grad_norm": 1.6525421142578125,
       "learning_rate": 6.948717948717949e-06,
-      "loss": 1.2489395141601562,
+      "loss": 1.2684237480163574,
       "step": 220
     },
     {
       "epoch": 0.9416581371545547,
-      "grad_norm": 2.6820783615112305,
+      "grad_norm": 4.144842624664307,
       "learning_rate": 6.692307692307692e-06,
-      "loss": 1.2914579391479493,
+      "loss": 1.3331901550292968,
       "step": 230
     },
     {
       "epoch": 0.9825997952917093,
-      "grad_norm": 4.043029308319092,
+      "grad_norm": 4.287148475646973,
       "learning_rate": 6.435897435897437e-06,
-      "loss": 1.2840445518493653,
+      "loss": 1.3455743789672852,
       "step": 240
     },
     {
       "epoch": 1.0204708290685773,
-      "grad_norm": 1.33917236328125,
+      "grad_norm": 1.7868481874465942,
       "learning_rate": 6.17948717948718e-06,
-      "loss": 1.20665283203125,
+      "loss": 1.2180957794189453,
       "step": 250
     },
     {
       "epoch": 1.0614124872057318,
-      "grad_norm": 3.203650712966919,
+      "grad_norm": 3.4123318195343018,
       "learning_rate": 5.923076923076924e-06,
-      "loss": 1.0558320999145507,
+      "loss": 1.0844655990600587,
       "step": 260
     },
     {
       "epoch": 1.1023541453428864,
-      "grad_norm": 2.068211078643799,
+      "grad_norm": 2.9107143878936768,
       "learning_rate": 5.666666666666667e-06,
-      "loss": 1.2426029205322267,
+      "loss": 1.2759632110595702,
       "step": 270
     },
     {
       "epoch": 1.143295803480041,
-      "grad_norm": 1.6644175052642822,
+      "grad_norm": 2.166071653366089,
       "learning_rate": 5.41025641025641e-06,
-      "loss": 1.2411796569824218,
+      "loss": 1.2902444839477538,
       "step": 280
     },
     {
       "epoch": 1.1842374616171956,
-      "grad_norm": 2.9029407501220703,
+      "grad_norm": 3.7286741733551025,
       "learning_rate": 5.1538461538461534e-06,
-      "loss": 1.0601702690124513,
+      "loss": 1.1132170677185058,
       "step": 290
     },
     {
       "epoch": 1.22517911975435,
-      "grad_norm": 2.7912986278533936,
+      "grad_norm": 3.644495964050293,
       "learning_rate": 4.8974358974358975e-06,
-      "loss": 1.2574929237365722,
+      "loss": 1.3006702423095704,
       "step": 300
     },
     {
       "epoch": 1.2661207778915047,
-      "grad_norm": 3.1852738857269287,
+      "grad_norm": 2.9774715900421143,
       "learning_rate": 4.641025641025642e-06,
-      "loss": 1.3007983207702636,
+      "loss": 1.3277738571166993,
       "step": 310
     },
     {
       "epoch": 1.3070624360286591,
-      "grad_norm": 2.374436616897583,
+      "grad_norm": 3.5428693294525146,
       "learning_rate": 4.384615384615385e-06,
-      "loss": 1.1643206596374511,
+      "loss": 1.2050466537475586,
       "step": 320
     },
     {
       "epoch": 1.3480040941658138,
-      "grad_norm": 2.4103164672851562,
+      "grad_norm": 2.9129176139831543,
       "learning_rate": 4.128205128205128e-06,
-      "loss": 1.0461477279663085,
+      "loss": 1.0705522537231444,
       "step": 330
     },
     {
       "epoch": 1.3889457523029682,
-      "grad_norm": 2.588911533355713,
+      "grad_norm": 3.6323633193969727,
       "learning_rate": 3.871794871794872e-06,
-      "loss": 1.152013111114502,
+      "loss": 1.1811490058898926,
       "step": 340
     },
     {
       "epoch": 1.429887410440123,
-      "grad_norm": 2.1230835914611816,
+      "grad_norm": 2.344156265258789,
       "learning_rate": 3.6153846153846156e-06,
-      "loss": 1.2414540290832519,
+      "loss": 1.2702527046203613,
       "step": 350
     },
     {
       "epoch": 1.4708290685772774,
-      "grad_norm": 2.0331356525421143,
+      "grad_norm": 2.93733286857605,
       "learning_rate": 3.358974358974359e-06,
-      "loss": 1.221489429473877,
+      "loss": 1.2678390502929688,
       "step": 360
     },
     {
       "epoch": 1.511770726714432,
-      "grad_norm": 2.5309641361236572,
+      "grad_norm": 3.3147552013397217,
       "learning_rate": 3.102564102564103e-06,
-      "loss": 0.9268945693969727,
+      "loss": 0.975819206237793,
       "step": 370
     },
     {
       "epoch": 1.5527123848515865,
-      "grad_norm": 2.3488223552703857,
+      "grad_norm": 3.5217061042785645,
       "learning_rate": 2.846153846153846e-06,
-      "loss": 1.5277278900146485,
+      "loss": 1.5488268852233886,
       "step": 380
     },
     {
       "epoch": 1.593654042988741,
-      "grad_norm": 2.196445941925049,
+      "grad_norm": 2.5279595851898193,
       "learning_rate": 2.5897435897435903e-06,
-      "loss": 1.1361653327941894,
+      "loss": 1.1411288261413575,
       "step": 390
     },
     {
       "epoch": 1.6345957011258956,
-      "grad_norm": 2.2993838787078857,
+      "grad_norm": 2.848078489303589,
       "learning_rate": 2.3333333333333336e-06,
-      "loss": 1.1851417541503906,
+      "loss": 1.1716268539428711,
       "step": 400
     }
   ],

sft/checkpoint-490/adapter_config.json

@@ -30,8 +30,8 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "q_proj",
     "k_proj",
+    "q_proj",
     "v_proj",
     "o_proj"
   ],

sft/checkpoint-490/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7d5efc1629290bcfa82fca8d6fe0a1fa199e6edd6673b12516e20f84ef8bf8f9
+oid sha256:f17065f015b60b2b71094531a596650f064bc0d4e2c5ae78ced1765b8daf35ac
 size 61380432

sft/checkpoint-490/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:35e3f008ba58149c4d3cf026a50e809c2dc5af6a08fc7f325152af82499501e0
+oid sha256:a3a60a8ba30916dbad0a22c74ab53e349e337af2bc29188447ad2e67d2854f41
 size 122930379

sft/checkpoint-490/trainer_state.json

@@ -11,345 +11,345 @@
   "log_history": [
     {
       "epoch": 0.04094165813715456,
-      "grad_norm": 6.233852386474609,
+      "grad_norm": 7.529000759124756,
       "learning_rate": 9.000000000000001e-07,
-      "loss": 6.206051254272461,
+      "loss": 6.639524841308594,
       "step": 10
     },
     {
       "epoch": 0.08188331627430911,
-      "grad_norm": 5.675178527832031,
+      "grad_norm": 7.1427388191223145,
       "learning_rate": 1.9000000000000002e-06,
-      "loss": 6.079831314086914,
+      "loss": 6.582770538330078,
       "step": 20
     },
     {
       "epoch": 0.12282497441146366,
-      "grad_norm": 5.132958889007568,
+      "grad_norm": 6.251121997833252,
       "learning_rate": 2.9e-06,
-      "loss": 5.749290466308594,
+      "loss": 6.114261245727539,
       "step": 30
     },
     {
       "epoch": 0.16376663254861823,
-      "grad_norm": 5.81948184967041,
+      "grad_norm": 7.014133453369141,
       "learning_rate": 3.900000000000001e-06,
-      "loss": 6.033080673217773,
+      "loss": 6.333962631225586,
       "step": 40
     },
     {
       "epoch": 0.2047082906857728,
-      "grad_norm": 6.442906379699707,
+      "grad_norm": 7.072969913482666,
       "learning_rate": 4.9000000000000005e-06,
-      "loss": 5.972381973266602,
+      "loss": 6.248543930053711,
       "step": 50
     },
     {
       "epoch": 0.24564994882292732,
-      "grad_norm": 5.304522514343262,
+      "grad_norm": 5.674424171447754,
       "learning_rate": 5.9e-06,
-      "loss": 5.472136306762695,
+      "loss": 5.6220745086669925,
       "step": 60
     },
     {
       "epoch": 0.2865916069600819,
-      "grad_norm": 5.261083126068115,
+      "grad_norm": 5.3739800453186035,
       "learning_rate": 6.9e-06,
-      "loss": 5.041688537597656,
+      "loss": 5.0251930236816404,
       "step": 70
     },
     {
       "epoch": 0.32753326509723646,
-      "grad_norm": 6.387202739715576,
+      "grad_norm": 6.372293472290039,
       "learning_rate": 7.9e-06,
-      "loss": 4.916830062866211,
+      "loss": 5.033017349243164,
       "step": 80
     },
     {
       "epoch": 0.368474923234391,
-      "grad_norm": 7.068334102630615,
+      "grad_norm": 7.22106409072876,
       "learning_rate": 8.900000000000001e-06,
-      "loss": 3.531498336791992,
+      "loss": 3.514456939697266,
       "step": 90
     },
     {
       "epoch": 0.4094165813715456,
-      "grad_norm": 4.787615776062012,
+      "grad_norm": 4.990958213806152,
       "learning_rate": 9.9e-06,
-      "loss": 2.2226877212524414,
+      "loss": 2.256962776184082,
       "step": 100
     },
     {
       "epoch": 0.4503582395087001,
-      "grad_norm": 3.4490268230438232,
+      "grad_norm": 3.2661993503570557,
       "learning_rate": 9.76923076923077e-06,
-      "loss": 1.955281639099121,
+      "loss": 2.0482555389404298,
       "step": 110
     },
     {
       "epoch": 0.49129989764585463,
-      "grad_norm": 2.435624361038208,
+      "grad_norm": 2.997323513031006,
       "learning_rate": 9.512820512820514e-06,
-      "loss": 1.5814408302307128,
+      "loss": 1.6077747344970703,
       "step": 120
     },
     {
       "epoch": 0.5322415557830092,
-      "grad_norm": 2.0870230197906494,
+      "grad_norm": 2.422741413116455,
       "learning_rate": 9.256410256410257e-06,
-      "loss": 1.4773900985717774,
+      "loss": 1.538028621673584,
       "step": 130
     },
     {
       "epoch": 0.5731832139201638,
-      "grad_norm": 2.1049163341522217,
+      "grad_norm": 2.4594290256500244,
       "learning_rate": 9e-06,
-      "loss": 1.508018970489502,
+      "loss": 1.5614049911499024,
       "step": 140
     },
     {
       "epoch": 0.6141248720573184,
-      "grad_norm": 2.7535016536712646,
+      "grad_norm": 2.7256577014923096,
       "learning_rate": 8.743589743589743e-06,
-      "loss": 1.6693695068359375,
+      "loss": 1.7180768966674804,
       "step": 150
     },
     {
       "epoch": 0.6550665301944729,
-      "grad_norm": 2.2710325717926025,
+      "grad_norm": 2.6614902019500732,
       "learning_rate": 8.487179487179488e-06,
-      "loss": 1.590962028503418,
+      "loss": 1.6412044525146485,
       "step": 160
     },
     {
       "epoch": 0.6960081883316275,
-      "grad_norm": 2.007903575897217,
+      "grad_norm": 2.1517934799194336,
       "learning_rate": 8.230769230769232e-06,
-      "loss": 1.6514751434326171,
+      "loss": 1.729467010498047,
       "step": 170
     },
     {
       "epoch": 0.736949846468782,
-      "grad_norm": 2.2933390140533447,
+      "grad_norm": 2.4588229656219482,
       "learning_rate": 7.974358974358975e-06,
-      "loss": 1.2684261322021484,
+      "loss": 1.305363941192627,
       "step": 180
     },
     {
       "epoch": 0.7778915046059366,
-      "grad_norm": 2.4877562522888184,
+      "grad_norm": 2.5282061100006104,
       "learning_rate": 7.717948717948718e-06,
-      "loss": 1.4245257377624512,
+      "loss": 1.4986873626708985,
       "step": 190
     },
     {
       "epoch": 0.8188331627430911,
-      "grad_norm": 2.4900834560394287,
+      "grad_norm": 3.699396848678589,
       "learning_rate": 7.461538461538462e-06,
-      "loss": 1.4417343139648438,
+      "loss": 1.5059691429138184,
       "step": 200
     },
     {
       "epoch": 0.8597748208802457,
-      "grad_norm": 1.8601328134536743,
+      "grad_norm": 2.2632241249084473,
       "learning_rate": 7.205128205128206e-06,
-      "loss": 1.3829060554504395,
+      "loss": 1.4249659538269044,
       "step": 210
     },
     {
       "epoch": 0.9007164790174002,
-      "grad_norm": 1.3816406726837158,
+      "grad_norm": 1.6525421142578125,
       "learning_rate": 6.948717948717949e-06,
-      "loss": 1.2489395141601562,
+      "loss": 1.2684237480163574,
       "step": 220
     },
     {
       "epoch": 0.9416581371545547,
-      "grad_norm": 2.6820783615112305,
+      "grad_norm": 4.144842624664307,
       "learning_rate": 6.692307692307692e-06,
-      "loss": 1.2914579391479493,
+      "loss": 1.3331901550292968,
       "step": 230
     },
     {
       "epoch": 0.9825997952917093,
-      "grad_norm": 4.043029308319092,
+      "grad_norm": 4.287148475646973,
       "learning_rate": 6.435897435897437e-06,
-      "loss": 1.2840445518493653,
+      "loss": 1.3455743789672852,
       "step": 240
     },
     {
       "epoch": 1.0204708290685773,
-      "grad_norm": 1.33917236328125,
+      "grad_norm": 1.7868481874465942,
       "learning_rate": 6.17948717948718e-06,
-      "loss": 1.20665283203125,
+      "loss": 1.2180957794189453,
       "step": 250
     },
     {
       "epoch": 1.0614124872057318,
-      "grad_norm": 3.203650712966919,
+      "grad_norm": 3.4123318195343018,
       "learning_rate": 5.923076923076924e-06,
-      "loss": 1.0558320999145507,
+      "loss": 1.0844655990600587,
       "step": 260
     },
     {
       "epoch": 1.1023541453428864,
-      "grad_norm": 2.068211078643799,
+      "grad_norm": 2.9107143878936768,
       "learning_rate": 5.666666666666667e-06,
-      "loss": 1.2426029205322267,
+      "loss": 1.2759632110595702,
       "step": 270
     },
     {
       "epoch": 1.143295803480041,
-      "grad_norm": 1.6644175052642822,
+      "grad_norm": 2.166071653366089,
       "learning_rate": 5.41025641025641e-06,
-      "loss": 1.2411796569824218,
+      "loss": 1.2902444839477538,
       "step": 280
     },
     {
       "epoch": 1.1842374616171956,
-      "grad_norm": 2.9029407501220703,
+      "grad_norm": 3.7286741733551025,
       "learning_rate": 5.1538461538461534e-06,
-      "loss": 1.0601702690124513,
+      "loss": 1.1132170677185058,
       "step": 290
     },
     {
       "epoch": 1.22517911975435,
-      "grad_norm": 2.7912986278533936,
+      "grad_norm": 3.644495964050293,
       "learning_rate": 4.8974358974358975e-06,
-      "loss": 1.2574929237365722,
+      "loss": 1.3006702423095704,
       "step": 300
     },
     {
       "epoch": 1.2661207778915047,
-      "grad_norm": 3.1852738857269287,
+      "grad_norm": 2.9774715900421143,
       "learning_rate": 4.641025641025642e-06,
-      "loss": 1.3007983207702636,
+      "loss": 1.3277738571166993,
       "step": 310
     },
     {
       "epoch": 1.3070624360286591,
-      "grad_norm": 2.374436616897583,
+      "grad_norm": 3.5428693294525146,
       "learning_rate": 4.384615384615385e-06,
-      "loss": 1.1643206596374511,
+      "loss": 1.2050466537475586,
       "step": 320
     },
     {
       "epoch": 1.3480040941658138,
-      "grad_norm": 2.4103164672851562,
+      "grad_norm": 2.9129176139831543,
       "learning_rate": 4.128205128205128e-06,
-      "loss": 1.0461477279663085,
+      "loss": 1.0705522537231444,
       "step": 330
     },
     {
       "epoch": 1.3889457523029682,
-      "grad_norm": 2.588911533355713,
+      "grad_norm": 3.6323633193969727,
       "learning_rate": 3.871794871794872e-06,
-      "loss": 1.152013111114502,
+      "loss": 1.1811490058898926,
       "step": 340
     },
     {
       "epoch": 1.429887410440123,
-      "grad_norm": 2.1230835914611816,
+      "grad_norm": 2.344156265258789,
       "learning_rate": 3.6153846153846156e-06,
-      "loss": 1.2414540290832519,
+      "loss": 1.2702527046203613,
       "step": 350
     },
     {
       "epoch": 1.4708290685772774,
-      "grad_norm": 2.0331356525421143,
+      "grad_norm": 2.93733286857605,
       "learning_rate": 3.358974358974359e-06,
-      "loss": 1.221489429473877,
+      "loss": 1.2678390502929688,
       "step": 360
     },
     {
       "epoch": 1.511770726714432,
-      "grad_norm": 2.5309641361236572,
+      "grad_norm": 3.3147552013397217,
       "learning_rate": 3.102564102564103e-06,
-      "loss": 0.9268945693969727,
+      "loss": 0.975819206237793,
       "step": 370
     },
     {
       "epoch": 1.5527123848515865,
-      "grad_norm": 2.3488223552703857,
+      "grad_norm": 3.5217061042785645,
       "learning_rate": 2.846153846153846e-06,
-      "loss": 1.5277278900146485,
+      "loss": 1.5488268852233886,
       "step": 380
     },
     {
       "epoch": 1.593654042988741,
-      "grad_norm": 2.196445941925049,
+      "grad_norm": 2.5279595851898193,
       "learning_rate": 2.5897435897435903e-06,
-      "loss": 1.1361653327941894,
+      "loss": 1.1411288261413575,
       "step": 390
     },
     {
       "epoch": 1.6345957011258956,
-      "grad_norm": 2.2993838787078857,
+      "grad_norm": 2.848078489303589,
       "learning_rate": 2.3333333333333336e-06,
-      "loss": 1.1851417541503906,
+      "loss": 1.1716268539428711,
       "step": 400
     },
     {
       "epoch": 1.6755373592630503,
-      "grad_norm": 1.7422572374343872,
+      "grad_norm": 2.8351571559906006,
       "learning_rate": 2.0769230769230773e-06,
-      "loss": 1.0863821983337403,
+      "loss": 1.139915370941162,
       "step": 410
     },
     {
       "epoch": 1.7164790174002047,
-      "grad_norm": 2.5725717544555664,
+      "grad_norm": 5.1406707763671875,
       "learning_rate": 1.8205128205128205e-06,
-      "loss": 0.9950428962707519,
+      "loss": 1.016810417175293,
       "step": 420
     },
     {
       "epoch": 1.7574206755373591,
-      "grad_norm": 2.113790988922119,
+      "grad_norm": 2.3416874408721924,
       "learning_rate": 1.5641025641025642e-06,
-      "loss": 0.9922480583190918,
+      "loss": 1.0270769119262695,
       "step": 430
     },
     {
       "epoch": 1.7983623336745138,
-      "grad_norm": 1.7282646894454956,
+      "grad_norm": 2.4181466102600098,
       "learning_rate": 1.307692307692308e-06,
-      "loss": 1.1919940948486327,
+      "loss": 1.2266251564025878,
       "step": 440
     },
     {
       "epoch": 1.8393039918116685,
-      "grad_norm": 2.1598236560821533,
+      "grad_norm": 2.907052755355835,
       "learning_rate": 1.0512820512820514e-06,
-      "loss": 1.173375701904297,
+      "loss": 1.1883393287658692,
       "step": 450
     },
     {
       "epoch": 1.880245649948823,
-      "grad_norm": 2.2551820278167725,
+      "grad_norm": 2.765097141265869,
       "learning_rate": 7.948717948717949e-07,
-      "loss": 1.1812771797180175,
+      "loss": 1.2259196281433105,
       "step": 460
     },
     {
       "epoch": 1.9211873080859774,
-      "grad_norm": 2.8052544593811035,
+      "grad_norm": 3.7463090419769287,
       "learning_rate": 5.384615384615386e-07,
-      "loss": 0.9614218711853028,
+      "loss": 1.016909694671631,
       "step": 470
     },
     {
       "epoch": 1.962128966223132,
-      "grad_norm": 3.274524450302124,
+      "grad_norm": 3.9370245933532715,
       "learning_rate": 2.820512820512821e-07,
-      "loss": 1.0670873641967773,
+      "loss": 1.1455986976623536,
       "step": 480
     },
     {
       "epoch": 2.0,
-      "grad_norm": 6.130195617675781,
+      "grad_norm": 4.772556304931641,
       "learning_rate": 2.5641025641025643e-08,
-      "loss": 1.051560401916504,
+      "loss": 1.0747014999389648,
       "step": 490
     }
   ],

sft/final/adapter_config.json

@@ -30,8 +30,8 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "q_proj",
     "k_proj",
+    "q_proj",
     "v_proj",
     "o_proj"
   ],

sft/final/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7d5efc1629290bcfa82fca8d6fe0a1fa199e6edd6673b12516e20f84ef8bf8f9
+oid sha256:f17065f015b60b2b71094531a596650f064bc0d4e2c5ae78ced1765b8daf35ac
 size 61380432