Upload folder using huggingface_hub

README.md

@@ -1,11 +1,12 @@
 ---
+library_name: transformers
 license: llama3
 base_model: meta-llama/Meta-Llama-3-8B-Instruct
 tags:
 - alignment-handbook
 - generated_from_trainer
 datasets:
-- simonycl/llama3-ultrafeedback-annotate-judge-5
+- simonycl/Llama-3-8B-Instruct-ultrafeedback-judge-5-annotate
 model-index:
 - name: llama-3-8b-instruct-agg-judge
   results: []
@@ -16,17 +17,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # llama-3-8b-instruct-agg-judge
 
-This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on the simonycl/llama3-ultrafeedback-annotate-judge-5 dataset.
+This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on the simonycl/Llama-3-8B-Instruct-ultrafeedback-judge-5-annotate dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.6314
-- Rewards/chosen: -1.4785
-- Rewards/rejected: -1.8808
-- Rewards/accuracies: 0.6402
-- Rewards/margins: 0.4022
-- Logps/rejected: -332.7695
-- Logps/chosen: -294.6159
-- Logits/rejected: -1.5412
-- Logits/chosen: -1.5384
+- Loss: 0.5049
+- Rewards/chosen: -1.4456
+- Rewards/rejected: -2.1220
+- Rewards/accuracies: 0.7933
+- Rewards/margins: 0.6764
+- Logps/rejected: -356.5251
+- Logps/chosen: -307.2105
+- Logits/rejected: -1.1659
+- Logits/chosen: -0.9705
 
 ## Model description
 
@@ -46,14 +47,14 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 5e-07
-- train_batch_size: 2
-- eval_batch_size: 4
+- train_batch_size: 1
+- eval_batch_size: 1
 - seed: 42
 - distributed_type: multi-GPU
 - num_devices: 4
-- gradient_accumulation_steps: 16
+- gradient_accumulation_steps: 32
 - total_train_batch_size: 128
-- total_eval_batch_size: 16
+- total_eval_batch_size: 4
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: cosine
 - lr_scheduler_warmup_ratio: 0.1
@@ -63,12 +64,12 @@ 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 |
 |:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
-| 0.5551        | 0.8550 | 400  | 0.6314          | -1.4785        | -1.8808          | 0.6402             | 0.4022          | -332.7695      | -294.6159    | -1.5412         | -1.5384       |
+| 0.5505        | 0.8528 | 400  | 0.5049          | -1.4456        | -2.1220          | 0.7933             | 0.6764          | -356.5251      | -307.2105    | -1.1659         | -0.9705       |
 
 
 ### Framework versions
 
-- Transformers 4.44.0
+- Transformers 4.44.2
 - Pytorch 2.4.0+cu121
 - Datasets 2.21.0
 - Tokenizers 0.19.1

all_results.json

@@ -1,22 +1,22 @@
 {
-    "epoch": 0.9982631930527722,
-    "eval_logits/chosen": -1.5078670978546143,
-    "eval_logits/rejected": -1.5112224817276,
-    "eval_logps/chosen": -295.1907958984375,
-    "eval_logps/rejected": -333.72509765625,
-    "eval_loss": 0.6313705444335938,
-    "eval_rewards/accuracies": 0.6361788511276245,
-    "eval_rewards/chosen": -1.484257698059082,
-    "eval_rewards/margins": 0.40605515241622925,
-    "eval_rewards/rejected": -1.890312671661377,
-    "eval_runtime": 269.2761,
-    "eval_samples": 1961,
-    "eval_samples_per_second": 7.282,
-    "eval_steps_per_second": 0.457,
+    "epoch": 0.9999333733093477,
+    "eval_logits/chosen": -0.9731917977333069,
+    "eval_logits/rejected": -1.166278600692749,
+    "eval_logps/chosen": -306.48046875,
+    "eval_logps/rejected": -356.2023010253906,
+    "eval_loss": 0.503829836845398,
+    "eval_rewards/accuracies": 0.7933906316757202,
+    "eval_rewards/chosen": -1.4382753372192383,
+    "eval_rewards/margins": 0.6804661750793457,
+    "eval_rewards/rejected": -2.118741512298584,
+    "eval_runtime": 11557.6821,
+    "eval_samples": 60035,
+    "eval_samples_per_second": 5.194,
+    "eval_steps_per_second": 1.299,
     "total_flos": 0.0,
-    "train_loss": 0.5881322287900544,
-    "train_runtime": 16866.5046,
-    "train_samples": 59875,
-    "train_samples_per_second": 3.55,
-    "train_steps_per_second": 0.028
+    "train_loss": 0.5891387982409138,
+    "train_runtime": 37343.5856,
+    "train_samples": 60035,
+    "train_samples_per_second": 1.608,
+    "train_steps_per_second": 0.013
 }

config.json

@@ -23,7 +23,7 @@
   "rope_theta": 500000.0,
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
-  "transformers_version": "4.44.0",
+  "transformers_version": "4.44.2",
   "use_cache": true,
   "vocab_size": 128256
 }

eval_results.json

@@ -1,16 +1,16 @@
 {
-    "epoch": 0.9982631930527722,
-    "eval_logits/chosen": -1.5078670978546143,
-    "eval_logits/rejected": -1.5112224817276,
-    "eval_logps/chosen": -295.1907958984375,
-    "eval_logps/rejected": -333.72509765625,
-    "eval_loss": 0.6313705444335938,
-    "eval_rewards/accuracies": 0.6361788511276245,
-    "eval_rewards/chosen": -1.484257698059082,
-    "eval_rewards/margins": 0.40605515241622925,
-    "eval_rewards/rejected": -1.890312671661377,
-    "eval_runtime": 269.2761,
-    "eval_samples": 1961,
-    "eval_samples_per_second": 7.282,
-    "eval_steps_per_second": 0.457
+    "epoch": 0.9999333733093477,
+    "eval_logits/chosen": -0.9731917977333069,
+    "eval_logits/rejected": -1.166278600692749,
+    "eval_logps/chosen": -306.48046875,
+    "eval_logps/rejected": -356.2023010253906,
+    "eval_loss": 0.503829836845398,
+    "eval_rewards/accuracies": 0.7933906316757202,
+    "eval_rewards/chosen": -1.4382753372192383,
+    "eval_rewards/margins": 0.6804661750793457,
+    "eval_rewards/rejected": -2.118741512298584,
+    "eval_runtime": 11557.6821,
+    "eval_samples": 60035,
+    "eval_samples_per_second": 5.194,
+    "eval_steps_per_second": 1.299
 }

generation_config.json

@@ -8,5 +8,5 @@
   "max_length": 4096,
   "temperature": 0.6,
   "top_p": 0.9,
-  "transformers_version": "4.44.0"
+  "transformers_version": "4.44.2"
 }

model-00001-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d59998e606338c227db713796a769dd05db4458be1be180128e612a003ddcf5e
+oid sha256:aadc0d089c01c51929020ecb50c2681dbe21c6798b2542a82bb3fb4a7391f490
 size 4976698672

model-00002-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:69ab9ce83956487448a3c54fa5c473cb4c1699e99b1eec6529a448bea9082790
+oid sha256:799708995ad93f68aa9f73c9fdd0382bf183da7bd6496f320b063ee28ce7f144
 size 4999802720

model-00003-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a282710b40678e35b23e4c01bd51d4997708c501941b4fdcaa015bdee2e823bc
+oid sha256:948d4df220f4c7139f9b504c8eec9bd66b959994c3cdf878bbff15e5dbbd62dd
 size 4915916176

model-00004-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:df32c2bf8a67fdf19919b66b28ed3a6478d8bde44defa62902bee27fdb5efdad
+oid sha256:74cc414c0293f182445de8a80f8c3ddd96095a8d563049d72584d4a715e6831d
 size 1168138808

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 0.9982631930527722,
+    "epoch": 0.9999333733093477,
     "total_flos": 0.0,
-    "train_loss": 0.5881322287900544,
-    "train_runtime": 16866.5046,
-    "train_samples": 59875,
-    "train_samples_per_second": 3.55,
-    "train_steps_per_second": 0.028
+    "train_loss": 0.5891387982409138,
+    "train_runtime": 37343.5856,
+    "train_samples": 60035,
+    "train_samples_per_second": 1.608,
+    "train_steps_per_second": 0.013
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fca683b1c8da8ae65b1c44ba6260b682c262d2b793b0a08fc2aaee2b7d0426bd
+oid sha256:9a6681a55e194ed88e0b0738936a1587e12b39b5d17f641aed90732330e377db
 size 7544