Training in progress, epoch 8

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cae542a3f3b360afc498f714131caf0826a28582dfda48acd40f67a5e1f762c3
+oid sha256:b3f750124e4d89a3026dcc378ca2ca45664b8ea7a69439f4d5bb6a5c9be6e7cd
 size 26235704

checkpoint-23705/README.md

@@ -4,6 +4,17 @@ library_name: peft
 ## Training procedure
 
 
+The following `bitsandbytes` quantization config was used during training:
+- load_in_8bit: False
+- load_in_4bit: True
+- llm_int8_threshold: 6.0
+- llm_int8_skip_modules: None
+- llm_int8_enable_fp32_cpu_offload: False
+- llm_int8_has_fp16_weight: False
+- bnb_4bit_quant_type: nf4
+- bnb_4bit_use_double_quant: True
+- bnb_4bit_compute_dtype: bfloat16
+
 The following `bitsandbytes` quantization config was used during training:
 - load_in_8bit: False
 - load_in_4bit: True
@@ -16,5 +27,6 @@ The following `bitsandbytes` quantization config was used during training:
 - bnb_4bit_compute_dtype: bfloat16
 ### Framework versions
 
+- PEFT 0.4.0
 
 - PEFT 0.4.0

{checkpoint-6773 → checkpoint-23705}/adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:838944a5b86a3272e4404377409a27a6e121a7cb84b1c8c92ee6f7b274d25d0b
+oid sha256:4d4ea0239dac0417bfc6b6d1a1ec55b59d2550d6de3ff6085cda4a4fdd52e71a
 size 26271757

{checkpoint-6773 → checkpoint-27092}/README.md

@@ -4,17 +4,6 @@ library_name: peft
 ## Training procedure
 
 
-The following `bitsandbytes` quantization config was used during training:
-- load_in_8bit: False
-- load_in_4bit: True
-- llm_int8_threshold: 6.0
-- llm_int8_skip_modules: None
-- llm_int8_enable_fp32_cpu_offload: False
-- llm_int8_has_fp16_weight: False
-- bnb_4bit_quant_type: nf4
-- bnb_4bit_use_double_quant: True
-- bnb_4bit_compute_dtype: bfloat16
-
 The following `bitsandbytes` quantization config was used during training:
 - load_in_8bit: False
 - load_in_4bit: True
@@ -27,6 +16,5 @@ The following `bitsandbytes` quantization config was used during training:
 - bnb_4bit_compute_dtype: bfloat16
 ### Framework versions
 
-- PEFT 0.4.0
 
 - PEFT 0.4.0

{checkpoint-6773 → checkpoint-27092}/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dc4795225c78aa16f43da67a25351858aa53a60e20d4db2258e8ff9f51649397
+oid sha256:b3f750124e4d89a3026dcc378ca2ca45664b8ea7a69439f4d5bb6a5c9be6e7cd
 size 26235704

{checkpoint-6773 → checkpoint-27092}/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9c1c4cb036c320cebd5f4f96a4d4e993ebd4e217517d858dc25bd71178a1c66d
+oid sha256:f23f42b2a24eba2c6516b96003d4f60c9ee2e147aba5cc1675e4632e2233ffa2
 size 52562821

{checkpoint-6773 → checkpoint-27092}/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e663be8ea93d417fe6e48a37b2e00dd1e7c5f0e13ad67f0084b36b12c356eff6
+oid sha256:42ec19c306ea3c525ecdf98721b7a6c55e495e3c5ae1a63e9043cc7fc4e158a2
 size 14575

{checkpoint-6773 → checkpoint-27092}/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8d0b9b1a1bf0a512b86fb2d695d77c4fb40483963c22a2f878ac8390632405e7
+oid sha256:84283001aaebc22c26d4a3f0bdb08995525d4e7773927a4ace9984e3cac9bf00
 size 627

checkpoint-27092/trainer_state.json

@@ -0,0 +1,404 @@
+{
+  "best_metric": 0.7249947190284729,
+  "best_model_checkpoint": "./Zeroshot/01-12-23-NousResearch-Nous-Hermes-Llama2-13b_multilang-dataset-3.0.3-portuguese-2_epochs-10_batch_2/checkpoints/checkpoint-20319",
+  "epoch": 8.0,
+  "global_step": 27092,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.8948611931482576e-05,
+      "loss": 1.3183,
+      "step": 500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00011801535735380982,
+      "loss": 0.9343,
+      "step": 1000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00017708210277613703,
+      "loss": 0.9005,
+      "step": 1500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.00023614884819846425,
+      "loss": 0.8859,
+      "step": 2000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 0.0002952155936207915,
+      "loss": 0.8707,
+      "step": 2500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 0.00035416420555227406,
+      "loss": 0.8523,
+      "step": 3000
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 0.8459745645523071,
+      "eval_runtime": 133.0397,
+      "eval_samples_per_second": 11.32,
+      "eval_steps_per_second": 1.421,
+      "step": 3386
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 0.00039998666871418663,
+      "loss": 0.8395,
+      "step": 3500
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 0.00039960207649127196,
+      "loss": 0.8052,
+      "step": 4000
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 0.0003986872720605752,
+      "loss": 0.8036,
+      "step": 4500
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 0.00039724468545932536,
+      "loss": 0.7902,
+      "step": 5000
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 0.0003952781486968895,
+      "loss": 0.7818,
+      "step": 5500
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 0.00039279288557562877,
+      "loss": 0.7834,
+      "step": 6000
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 0.0003897954978146717,
+      "loss": 0.7686,
+      "step": 6500
+    },
+    {
+      "epoch": 2.0,
+      "eval_loss": 0.7881951332092285,
+      "eval_runtime": 132.7462,
+      "eval_samples_per_second": 11.345,
+      "eval_steps_per_second": 1.424,
+      "step": 6773
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 0.0003862939475134625,
+      "loss": 0.7371,
+      "step": 7000
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 0.00038230601585633047,
+      "loss": 0.6893,
+      "step": 7500
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 0.0003778263160223725,
+      "loss": 0.6936,
+      "step": 8000
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 0.00037287424793815686,
+      "loss": 0.6923,
+      "step": 8500
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 0.0003674629660106825,
+      "loss": 0.6862,
+      "step": 9000
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 0.00036160684447787385,
+      "loss": 0.6869,
+      "step": 9500
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 0.0003553214392256293,
+      "loss": 0.6856,
+      "step": 10000
+    },
+    {
+      "epoch": 3.0,
+      "eval_loss": 0.7562845945358276,
+      "eval_runtime": 132.7178,
+      "eval_samples_per_second": 11.347,
+      "eval_steps_per_second": 1.424,
+      "step": 10159
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 0.0003486234464660031,
+      "loss": 0.6199,
+      "step": 10500
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 0.0003415452256149365,
+      "loss": 0.595,
+      "step": 11000
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 0.0003340772155486902,
+      "loss": 0.5962,
+      "step": 11500
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 0.0003262690400990177,
+      "loss": 0.6019,
+      "step": 12000
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 0.00031811015213363493,
+      "loss": 0.6043,
+      "step": 12500
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 0.0003096375227201415,
+      "loss": 0.5984,
+      "step": 13000
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 0.0003008736580950757,
+      "loss": 0.6011,
+      "step": 13500
+    },
+    {
+      "epoch": 4.0,
+      "eval_loss": 0.738968551158905,
+      "eval_runtime": 132.6823,
+      "eval_samples_per_second": 11.35,
+      "eval_steps_per_second": 1.424,
+      "step": 13546
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 0.00029184183811650945,
+      "loss": 0.5125,
+      "step": 14000
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 0.00028256605442468665,
+      "loss": 0.5143,
+      "step": 14500
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 0.00027307094671192217,
+      "loss": 0.5171,
+      "step": 15000
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 0.0002633817372710499,
+      "loss": 0.5133,
+      "step": 15500
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 0.00025352416399628225,
+      "loss": 0.5155,
+      "step": 16000
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 0.0002435244120144539,
+      "loss": 0.5217,
+      "step": 16500
+    },
+    {
+      "epoch": 5.0,
+      "eval_loss": 0.7306132912635803,
+      "eval_runtime": 132.674,
+      "eval_samples_per_second": 11.351,
+      "eval_steps_per_second": 1.425,
+      "step": 16932
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 0.00023342937243316328,
+      "loss": 0.5015,
+      "step": 17000
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 0.00022322540906831537,
+      "loss": 0.4297,
+      "step": 17500
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 0.00021295975097746452,
+      "loss": 0.4336,
+      "step": 18000
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 0.00020265966730197618,
+      "loss": 0.4382,
+      "step": 18500
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 0.00019235251862948508,
+      "loss": 0.4357,
+      "step": 19000
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 0.0001820862195310506,
+      "loss": 0.4411,
+      "step": 19500
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 0.0001718469024885363,
+      "loss": 0.4411,
+      "step": 20000
+    },
+    {
+      "epoch": 6.0,
+      "eval_loss": 0.7249947190284729,
+      "eval_runtime": 132.741,
+      "eval_samples_per_second": 11.345,
+      "eval_steps_per_second": 1.424,
+      "step": 20319
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 0.00016168236981825817,
+      "loss": 0.4086,
+      "step": 20500
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 0.00015161962203746017,
+      "loss": 0.3539,
+      "step": 21000
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 0.00014170511189134913,
+      "loss": 0.3628,
+      "step": 21500
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 0.00013192544705149724,
+      "loss": 0.3737,
+      "step": 22000
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 0.00012232661178840682,
+      "loss": 0.3739,
+      "step": 22500
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 0.00011293410393119082,
+      "loss": 0.3781,
+      "step": 23000
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 0.00010377287323195157,
+      "loss": 0.3773,
+      "step": 23500
+    },
+    {
+      "epoch": 7.0,
+      "eval_loss": 0.7291969656944275,
+      "eval_runtime": 132.6924,
+      "eval_samples_per_second": 11.35,
+      "eval_steps_per_second": 1.424,
+      "step": 23705
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 9.486725509060299e-05,
+      "loss": 0.3353,
+      "step": 24000
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 8.624090591162438e-05,
+      "loss": 0.3143,
+      "step": 24500
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 7.79167402644616e-05,
+      "loss": 0.3206,
+      "step": 25000
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 6.991687001449847e-05,
+      "loss": 0.3215,
+      "step": 25500
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 6.226254558628854e-05,
+      "loss": 0.317,
+      "step": 26000
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 5.497409951507149e-05,
+      "loss": 0.3158,
+      "step": 26500
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 4.807089243652132e-05,
+      "loss": 0.3103,
+      "step": 27000
+    },
+    {
+      "epoch": 8.0,
+      "eval_loss": 0.7368675470352173,
+      "eval_runtime": 132.8691,
+      "eval_samples_per_second": 11.334,
+      "eval_steps_per_second": 1.422,
+      "step": 27092
+    }
+  ],
+  "max_steps": 33860,
+  "num_train_epochs": 10,
+  "total_flos": 1.207084111578194e+18,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-6773/trainer_state.json

@@ -1,110 +0,0 @@
-{
-  "best_metric": 0.7881951332092285,
-  "best_model_checkpoint": "./Zeroshot/01-12-23-NousResearch-Nous-Hermes-Llama2-13b_multilang-dataset-3.0.3-portuguese-2_epochs-10_batch_2/checkpoints/checkpoint-6773",
-  "epoch": 2.0,
-  "global_step": 6773,
-  "is_hyper_param_search": false,
-  "is_local_process_zero": true,
-  "is_world_process_zero": true,
-  "log_history": [
-    {
-      "epoch": 0.15,
-      "learning_rate": 5.8948611931482576e-05,
-      "loss": 1.3183,
-      "step": 500
-    },
-    {
-      "epoch": 0.3,
-      "learning_rate": 0.00011801535735380982,
-      "loss": 0.9343,
-      "step": 1000
-    },
-    {
-      "epoch": 0.44,
-      "learning_rate": 0.00017708210277613703,
-      "loss": 0.9005,
-      "step": 1500
-    },
-    {
-      "epoch": 0.59,
-      "learning_rate": 0.00023614884819846425,
-      "loss": 0.8859,
-      "step": 2000
-    },
-    {
-      "epoch": 0.74,
-      "learning_rate": 0.0002952155936207915,
-      "loss": 0.8707,
-      "step": 2500
-    },
-    {
-      "epoch": 0.89,
-      "learning_rate": 0.00035416420555227406,
-      "loss": 0.8523,
-      "step": 3000
-    },
-    {
-      "epoch": 1.0,
-      "eval_loss": 0.8459745645523071,
-      "eval_runtime": 133.0397,
-      "eval_samples_per_second": 11.32,
-      "eval_steps_per_second": 1.421,
-      "step": 3386
-    },
-    {
-      "epoch": 1.03,
-      "learning_rate": 0.00039998666871418663,
-      "loss": 0.8395,
-      "step": 3500
-    },
-    {
-      "epoch": 1.18,
-      "learning_rate": 0.00039960207649127196,
-      "loss": 0.8052,
-      "step": 4000
-    },
-    {
-      "epoch": 1.33,
-      "learning_rate": 0.0003986872720605752,
-      "loss": 0.8036,
-      "step": 4500
-    },
-    {
-      "epoch": 1.48,
-      "learning_rate": 0.00039724468545932536,
-      "loss": 0.7902,
-      "step": 5000
-    },
-    {
-      "epoch": 1.62,
-      "learning_rate": 0.0003952781486968895,
-      "loss": 0.7818,
-      "step": 5500
-    },
-    {
-      "epoch": 1.77,
-      "learning_rate": 0.00039279288557562877,
-      "loss": 0.7834,
-      "step": 6000
-    },
-    {
-      "epoch": 1.92,
-      "learning_rate": 0.0003897954978146717,
-      "loss": 0.7686,
-      "step": 6500
-    },
-    {
-      "epoch": 2.0,
-      "eval_loss": 0.7881951332092285,
-      "eval_runtime": 132.7462,
-      "eval_samples_per_second": 11.345,
-      "eval_steps_per_second": 1.424,
-      "step": 6773
-    }
-  ],
-  "max_steps": 33860,
-  "num_train_epochs": 10,
-  "total_flos": 3.0202162838237184e+17,
-  "trial_name": null,
-  "trial_params": null
-}

runs/Dec01_03-27-22_baa8ebef2c59/events.out.tfevents.1701401247.baa8ebef2c59.118280.1

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:76ba73b76339f3a725fd68918dd3e1daf6c9d67739a11479f5d34831c874789c
-size 14100
+oid sha256:e8771a09a819a481eebc55187f5f267890f66f8f6630892f21cd4dcdf6cfed59
+size 15496