Model save

README.md

@@ -15,15 +15,15 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.5657
-- Rewards/chosen: -0.0834
-- Rewards/rejected: -0.5329
-- Rewards/accuracies: 0.7020
-- Rewards/margins: 0.4496
-- Logps/rejected: -224.6124
-- Logps/chosen: -265.4659
-- Logits/rejected: -2.0001
-- Logits/chosen: -2.1212
+- Loss: 0.5256
+- Rewards/chosen: -0.1539
+- Rewards/rejected: -0.9025
+- Rewards/accuracies: 0.7420
+- Rewards/margins: 0.7486
+- Logps/rejected: -228.3078
+- Logps/chosen: -266.1707
+- Logits/rejected: -1.9406
+- Logits/chosen: -2.0654
 
 ## Model description
 
@@ -48,8 +48,8 @@ The following hyperparameters were used during training:
 - seed: 42
 - distributed_type: multi-GPU
 - num_devices: 4
-- gradient_accumulation_steps: 32
-- total_train_batch_size: 256
+- gradient_accumulation_steps: 8
+- total_train_batch_size: 64
 - total_eval_batch_size: 16
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
@@ -60,9 +60,9 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
 |:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
-| 0.6123        | 1.0   | 242  | 0.6065          | 0.0003         | -0.2418          | 0.6660             | 0.2421          | -221.7009      | -264.6292    | -2.0235         | -2.1436       |
-| 0.5841        | 2.0   | 484  | 0.5733          | -0.0606        | -0.4698          | 0.7040             | 0.4092          | -223.9813      | -265.2384    | -2.0064         | -2.1272       |
-| 0.567         | 3.0   | 726  | 0.5657          | -0.0834        | -0.5329          | 0.7020             | 0.4496          | -224.6124      | -265.4659    | -2.0001         | -2.1212       |
+| 0.5516        | 1.0   | 968  | 0.5547          | -0.1140        | -0.6431          | 0.7160             | 0.5291          | -225.7137      | -265.7718    | -1.9903         | -2.1116       |
+| 0.5443        | 2.0   | 1936 | 0.5307          | -0.1506        | -0.8643          | 0.7420             | 0.7136          | -227.9256      | -266.1383    | -1.9496         | -2.0740       |
+| 0.5439        | 3.0   | 2904 | 0.5256          | -0.1539        | -0.9025          | 0.7420             | 0.7486          | -228.3078      | -266.1707    | -1.9406         | -2.0654       |
 
 
 ### Framework versions

adapter_config.json

@@ -16,10 +16,10 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "v_proj",
+    "q_proj",
     "o_proj",
-    "k_proj",
-    "q_proj"
+    "v_proj",
+    "k_proj"
   ],
   "task_type": "CAUSAL_LM"
 }

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3e7a5241ce92e10bd13a1420f6b4bcf5bc57f917f958a8d5fedf2c6ac50f9e09
+oid sha256:1186b42eacfe8d1486235be62a0cb61e38eb9d4ad390ebc2fa0f525c81c15379
 size 218138576

all_results.json

@@ -1,21 +1,21 @@
 {
     "epoch": 3.0,
-    "eval_logits/chosen": -2.1211588382720947,
-    "eval_logits/rejected": -2.000145435333252,
-    "eval_logps/chosen": -265.46588134765625,
-    "eval_logps/rejected": -224.6123809814453,
-    "eval_loss": 0.5657259225845337,
-    "eval_rewards/accuracies": 0.7020000219345093,
-    "eval_rewards/chosen": -0.08337792754173279,
-    "eval_rewards/margins": 0.44957080483436584,
-    "eval_rewards/rejected": -0.5329487323760986,
-    "eval_runtime": 600.8868,
+    "eval_logits/chosen": -2.065433979034424,
+    "eval_logits/rejected": -1.9405803680419922,
+    "eval_logps/chosen": -266.1706848144531,
+    "eval_logps/rejected": -228.30780029296875,
+    "eval_loss": 0.5255534052848816,
+    "eval_rewards/accuracies": 0.7419999837875366,
+    "eval_rewards/chosen": -0.15385985374450684,
+    "eval_rewards/margins": 0.7486297488212585,
+    "eval_rewards/rejected": -0.9024895429611206,
+    "eval_runtime": 601.6531,
     "eval_samples": 2000,
-    "eval_samples_per_second": 3.328,
+    "eval_samples_per_second": 3.324,
     "eval_steps_per_second": 0.208,
-    "train_loss": 0.6034470564241908,
-    "train_runtime": 84521.5132,
+    "train_loss": 0.5642068754707158,
+    "train_runtime": 89225.6094,
     "train_samples": 61966,
-    "train_samples_per_second": 2.199,
-    "train_steps_per_second": 0.009
+    "train_samples_per_second": 2.083,
+    "train_steps_per_second": 0.033
 }

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 3.0,
-    "eval_logits/chosen": -2.1211588382720947,
-    "eval_logits/rejected": -2.000145435333252,
-    "eval_logps/chosen": -265.46588134765625,
-    "eval_logps/rejected": -224.6123809814453,
-    "eval_loss": 0.5657259225845337,
-    "eval_rewards/accuracies": 0.7020000219345093,
-    "eval_rewards/chosen": -0.08337792754173279,
-    "eval_rewards/margins": 0.44957080483436584,
-    "eval_rewards/rejected": -0.5329487323760986,
-    "eval_runtime": 600.8868,
+    "eval_logits/chosen": -2.065433979034424,
+    "eval_logits/rejected": -1.9405803680419922,
+    "eval_logps/chosen": -266.1706848144531,
+    "eval_logps/rejected": -228.30780029296875,
+    "eval_loss": 0.5255534052848816,
+    "eval_rewards/accuracies": 0.7419999837875366,
+    "eval_rewards/chosen": -0.15385985374450684,
+    "eval_rewards/margins": 0.7486297488212585,
+    "eval_rewards/rejected": -0.9024895429611206,
+    "eval_runtime": 601.6531,
     "eval_samples": 2000,
-    "eval_samples_per_second": 3.328,
+    "eval_samples_per_second": 3.324,
     "eval_steps_per_second": 0.208
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 3.0,
-    "train_loss": 0.6034470564241908,
-    "train_runtime": 84521.5132,
+    "train_loss": 0.5642068754707158,
+    "train_runtime": 89225.6094,
     "train_samples": 61966,
-    "train_samples_per_second": 2.199,
-    "train_steps_per_second": 0.009
+    "train_samples_per_second": 2.083,
+    "train_steps_per_second": 0.033
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2cc2d01d777cefbd752ab7c173c33376515874ab7b2ff1d04da2abc8678e0b2d
+oid sha256:874c7dfdeea022eb928eb034857e76c997eed7862cb1023dd73adc365f621759
 size 4283