End of training

README.md

@@ -2,9 +2,15 @@
 license: llama2
 base_model: llama-2-nl/Llama-2-7b-hf-lora-original-sft
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- BramVanroy/ultra_feedback_dutch
 model-index:
 - name: Llama-2-7b-hf-lora-original-it
   results: []
@@ -15,17 +21,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # Llama-2-7b-hf-lora-original-it
 
-This model is a fine-tuned version of [llama-2-nl/Llama-2-7b-hf-lora-original-sft](https://huggingface.co/llama-2-nl/Llama-2-7b-hf-lora-original-sft) on an unknown dataset.
+This model is a fine-tuned version of [llama-2-nl/Llama-2-7b-hf-lora-original-sft](https://huggingface.co/llama-2-nl/Llama-2-7b-hf-lora-original-sft) on the BramVanroy/ultra_feedback_dutch dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3540
-- Rewards/chosen: 0.1166
-- Rewards/rejected: -0.9259
+- Loss: 0.3536
+- Rewards/chosen: 0.1143
+- Rewards/rejected: -0.9295
 - Rewards/accuracies: 0.9396
-- Rewards/margins: 1.0425
-- Logps/rejected: -547.4224
-- Logps/chosen: -600.8121
-- Logits/rejected: -0.8736
-- Logits/chosen: -0.9600
+- Rewards/margins: 1.0437
+- Logps/rejected: -547.4578
+- Logps/chosen: -600.8353
+- Logits/rejected: -0.8732
+- Logits/chosen: -0.9594
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 0.9990049751243781,
+    "eval_logits/chosen": -0.9594090580940247,
+    "eval_logits/rejected": -0.8731673955917358,
+    "eval_logps/chosen": -600.8353271484375,
+    "eval_logps/rejected": -547.457763671875,
+    "eval_loss": 0.3535589277744293,
+    "eval_rewards/accuracies": 0.9395522475242615,
+    "eval_rewards/chosen": 0.11425414681434631,
+    "eval_rewards/margins": 1.0437376499176025,
+    "eval_rewards/rejected": -0.9294834136962891,
+    "eval_runtime": 710.4378,
+    "eval_samples": 5359,
+    "eval_samples_per_second": 7.543,
+    "eval_steps_per_second": 0.472,
     "total_flos": 0.0,
     "train_loss": 0.4291752041731856,
     "train_runtime": 22849.6211,

config.json

@@ -0,0 +1,29 @@
+{
+  "_name_or_path": "llama-2-nl/Llama-2-7b-hf-lora-original-sft",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "max_position_embeddings": 4096,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 32,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": null,
+  "rope_theta": 10000.0,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.40.1",
+  "use_cache": true,
+  "vocab_size": 32000
+}

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 0.9990049751243781,
+    "eval_logits/chosen": -0.9594090580940247,
+    "eval_logits/rejected": -0.8731673955917358,
+    "eval_logps/chosen": -600.8353271484375,
+    "eval_logps/rejected": -547.457763671875,
+    "eval_loss": 0.3535589277744293,
+    "eval_rewards/accuracies": 0.9395522475242615,
+    "eval_rewards/chosen": 0.11425414681434631,
+    "eval_rewards/margins": 1.0437376499176025,
+    "eval_rewards/rejected": -0.9294834136962891,
+    "eval_runtime": 710.4378,
+    "eval_samples": 5359,
+    "eval_samples_per_second": 7.543,
+    "eval_steps_per_second": 0.472
+}

runs/May20_00-51-42_gpu530.dodrio.os/events.out.tfevents.1716183684.gpu530.dodrio.os.1926607.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4c3ff428614af664a805e625ed6686bab8301c946ce831c3db6c573ce209f4a5
+size 828