TheBloke
/

DPOpenHermes-7B-GGUF

@@ -346,13 +346,15 @@ And thank you again to a16z for their generous grant.
 ## OpenHermes x Notus x Neural
 This is an RL fine tuned model of [Teknium](https://huggingface.co/teknium)'s [OpenHermes-2.5-Mistral-7B](https://huggingface.co/teknium/OpenHermes-2.5-Mistral-7B) using the [Intel/orca_dpo_pairs](https://huggingface.co/datasets/Intel/orca_dpo_pairs) and [argilla/ultrafeedback-binarized-preferences](https://huggingface.co/datasets/argilla/ultrafeedback-binarized-preferences) preference datasets for reinforcement learning using Direct Preference Optimization (DPO)
 DPOpenHermes is trained using qLoRA. The adapter is also provided in this model repo.
-# Training Details
-[<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
 DPOpenHermes was trained on a single H100 80GB hosted on RunPod for ~10h for 0.6 epochs of the dataset.
@@ -408,84 +410,82 @@ In LM-Studio, simply select the ChatML Prefix on the settings side pane:
 ```
 |             Task             |Version| Metric |Value |   |Stderr|
 |------------------------------|------:|--------|-----:|---|-----:|
-|agieval_aqua_rat              |      0|acc     |0.2480|_  |0.0272|
-|                              |       |acc_norm|0.2520|_  |0.0273|
-|agieval_logiqa_en             |      0|acc     |0.3810|_  |0.0190|
-|                              |       |acc_norm|0.3856|_  |0.0191|
-|agieval_lsat_ar               |      0|acc     |0.2348|_  |0.0280|
-|                              |       |acc_norm|0.2304|_  |0.0278|
-|agieval_lsat_lr               |      0|acc     |0.5118|_  |0.0222|
-|                              |       |acc_norm|0.5196|_  |0.0221|
 |agieval_lsat_rc               |      0|acc     |0.5948|_  |0.0300|
-|                              |       |acc_norm|0.5688|_  |0.0303|
-|agieval_sat_en                |      0|acc     |0.7427|_  |0.0305|
-|                              |       |acc_norm|0.7427|_  |0.0305|
-|agieval_sat_en_without_passage|      0|acc     |0.4563|_  |0.0348|
-|                              |       |acc_norm|0.4515|_  |0.0348|
-|agieval_sat_math              |      0|acc     |0.3818|_  |0.0328|
-|                              |       |acc_norm|0.3682|_  |0.0326|
 ```
-Average: 0.4399
 ## BigBench Hard
 ```
-hf-causal-experimental (pretrained=openaccess-ai-collective/dpopenhermes-alpha-v1,dtype=bfloat16,trust_remote_code=True,use_accelerate=True), limit: None, provide_description: False, num_fewshot: 0, batch_size: 16
 |                      Task                      |Version|       Metric        |Value |   |Stderr|
 |------------------------------------------------|------:|---------------------|-----:|---|-----:|
-|bigbench_causal_judgement                       |      0|multiple_choice_grade|0.5632|_  |0.0361|
-|bigbench_date_understanding                     |      0|multiple_choice_grade|0.6612|_  |0.0247|
 |bigbench_disambiguation_qa                      |      0|multiple_choice_grade|0.3566|_  |0.0299|
 |bigbench_geometric_shapes                       |      0|multiple_choice_grade|0.2006|_  |0.0212|
-|                                                |       |exact_str_match      |0.0334|_  |0.0095|
-|bigbench_logical_deduction_five_objects         |      0|multiple_choice_grade|0.3020|_  |0.0206|
-|bigbench_logical_deduction_seven_objects        |      0|multiple_choice_grade|0.2086|_  |0.0154|
-|bigbench_logical_deduction_three_objects        |      0|multiple_choice_grade|0.5033|_  |0.0289|
-|bigbench_movie_recommendation                   |      0|multiple_choice_grade|0.4220|_  |0.0221|
 |bigbench_navigate                               |      0|multiple_choice_grade|0.5000|_  |0.0158|
-|bigbench_reasoning_about_colored_objects        |      0|multiple_choice_grade|0.7035|_  |0.0102|
-|bigbench_ruin_names                             |      0|multiple_choice_grade|0.4107|_  |0.0233|
-|bigbench_salient_translation_error_detection    |      0|multiple_choice_grade|0.2154|_  |0.0130|
-|bigbench_snarks                                 |      0|multiple_choice_grade|0.7127|_  |0.0337|
-|bigbench_sports_understanding                   |      0|multiple_choice_grade|0.6988|_  |0.0146|
-|bigbench_temporal_sequences                     |      0|multiple_choice_grade|0.4670|_  |0.0158|
-|bigbench_tracking_shuffled_objects_five_objects |      0|multiple_choice_grade|0.2072|_  |0.0115|
-|bigbench_tracking_shuffled_objects_seven_objects|      0|multiple_choice_grade|0.1731|_  |0.0090|
-|bigbench_tracking_shuffled_objects_three_objects|      0|multiple_choice_grade|0.5033|_  |0.0289|
 ```
-Average: 0.4338
 ## GPT4All
 ```
 |    Task     |Version| Metric |Value |   |Stderr|
 |-------------|------:|--------|-----:|---|-----:|
-|arc_challenge|      0|acc     |0.5930|_  |0.0144|
-|             |       |acc_norm|0.6323|_  |0.0141|
-|arc_easy     |      0|acc     |0.8443|_  |0.0074|
-|             |       |acc_norm|0.8295|_  |0.0077|
 |boolq        |      1|acc     |0.8599|_  |0.0061|
-|hellaswag    |      0|acc     |0.6548|_  |0.0047|
-|             |       |acc_norm|0.8365|_  |0.0037|
-|openbookqa   |      0|acc     |0.3520|_  |0.0214|
-|             |       |acc_norm|0.4640|_  |0.0223|
-|piqa         |      0|acc     |0.8210|_  |0.0089|
-|             |       |acc_norm|0.8335|_  |0.0087|
-|winogrande   |      0|acc     |0.7466|_  |0.0122|
 ```
-Average: 0.7431
 ## TruthfulQA
 ```
-hf-causal-experimental (pretrained=openaccess-ai-collective/dpopenhermes-alpha-v1,dtype=bfloat16,trust_remote_code=True,use_accelerate=True), limit: None, provide_description: False, num_fewshot: 0, batch_size: 16
 |    Task     |Version|Metric|Value |   |Stderr|
 |-------------|------:|------|-----:|---|-----:|
-|truthfulqa_mc|      1|mc1   |0.4186|_  |0.0173|
-|             |       |mc2   |0.5847|_  |0.0153|
 ```
 <!-- original-model-card end -->

 ## OpenHermes x Notus x Neural
+[<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
 This is an RL fine tuned model of [Teknium](https://huggingface.co/teknium)'s [OpenHermes-2.5-Mistral-7B](https://huggingface.co/teknium/OpenHermes-2.5-Mistral-7B) using the [Intel/orca_dpo_pairs](https://huggingface.co/datasets/Intel/orca_dpo_pairs) and [argilla/ultrafeedback-binarized-preferences](https://huggingface.co/datasets/argilla/ultrafeedback-binarized-preferences) preference datasets for reinforcement learning using Direct Preference Optimization (DPO)
 DPOpenHermes is trained using qLoRA. The adapter is also provided in this model repo.
+Errata: Due to an issue with the DPO-only version failing to generate an eos token, this model was additional SFT with 7000 rows from the openhermes dataset to teach the model to use the eos_token again to end the turn. This resulted in lower benchmark scores. You can find the original DPO-only model in the `dpo-v0` branch.
+# Training Details
 DPOpenHermes was trained on a single H100 80GB hosted on RunPod for ~10h for 0.6 epochs of the dataset.
 ```
 |             Task             |Version| Metric |Value |   |Stderr|
 |------------------------------|------:|--------|-----:|---|-----:|
+|agieval_aqua_rat              |      0|acc     |0.2559|_  |0.0274|
+|                              |       |acc_norm|0.2598|_  |0.0276|
+|agieval_logiqa_en             |      0|acc     |0.3733|_  |0.0190|
+|                              |       |acc_norm|0.3886|_  |0.0191|
+|agieval_lsat_ar               |      0|acc     |0.2522|_  |0.0287|
+|                              |       |acc_norm|0.2522|_  |0.0287|
+|agieval_lsat_lr               |      0|acc     |0.5137|_  |0.0222|
+|                              |       |acc_norm|0.5294|_  |0.0221|
 |agieval_lsat_rc               |      0|acc     |0.5948|_  |0.0300|
+|                              |       |acc_norm|0.5725|_  |0.0302|
+|agieval_sat_en                |      0|acc     |0.7379|_  |0.0307|
+|                              |       |acc_norm|0.7282|_  |0.0311|
+|agieval_sat_en_without_passage|      0|acc     |0.4466|_  |0.0347|
+|                              |       |acc_norm|0.4466|_  |0.0347|
+|agieval_sat_math              |      0|acc     |0.3909|_  |0.0330|
+|                              |       |acc_norm|0.3591|_  |0.0324|
 ```
+Average: 0.4364
 ## BigBench Hard
 ```
 |                      Task                      |Version|       Metric        |Value |   |Stderr|
 |------------------------------------------------|------:|---------------------|-----:|---|-----:|
+|bigbench_causal_judgement                       |      0|multiple_choice_grade|0.5684|_  |0.0360|
+|bigbench_date_understanding                     |      0|multiple_choice_grade|0.6667|_  |0.0246|
 |bigbench_disambiguation_qa                      |      0|multiple_choice_grade|0.3566|_  |0.0299|
 |bigbench_geometric_shapes                       |      0|multiple_choice_grade|0.2006|_  |0.0212|
+|                                                |       |exact_str_match      |0.0724|_  |0.0137|
+|bigbench_logical_deduction_five_objects         |      0|multiple_choice_grade|0.2980|_  |0.0205|
+|bigbench_logical_deduction_seven_objects        |      0|multiple_choice_grade|0.2071|_  |0.0153|
+|bigbench_logical_deduction_three_objects        |      0|multiple_choice_grade|0.5067|_  |0.0289|
+|bigbench_movie_recommendation                   |      0|multiple_choice_grade|0.4140|_  |0.0220|
 |bigbench_navigate                               |      0|multiple_choice_grade|0.5000|_  |0.0158|
+|bigbench_reasoning_about_colored_objects        |      0|multiple_choice_grade|0.6980|_  |0.0103|
+|bigbench_ruin_names                             |      0|multiple_choice_grade|0.4174|_  |0.0233|
+|bigbench_salient_translation_error_detection    |      0|multiple_choice_grade|0.2044|_  |0.0128|
+|bigbench_snarks                                 |      0|multiple_choice_grade|0.7238|_  |0.0333|
+|bigbench_sports_understanding                   |      0|multiple_choice_grade|0.6876|_  |0.0148|
+|bigbench_temporal_sequences                     |      0|multiple_choice_grade|0.4360|_  |0.0157|
+|bigbench_tracking_shuffled_objects_five_objects |      0|multiple_choice_grade|0.2112|_  |0.0115|
+|bigbench_tracking_shuffled_objects_seven_objects|      0|multiple_choice_grade|0.1754|_  |0.0091|
+|bigbench_tracking_shuffled_objects_three_objects|      0|multiple_choice_grade|0.5067|_  |0.0289|
 ```
+Average: 0.4321
 ## GPT4All
 ```
 |    Task     |Version| Metric |Value |   |Stderr|
 |-------------|------:|--------|-----:|---|-----:|
+|arc_challenge|      0|acc     |0.5862|_  |0.0144|
+|             |       |acc_norm|0.6297|_  |0.0141|
+|arc_easy     |      0|acc     |0.8472|_  |0.0074|
+|             |       |acc_norm|0.8321|_  |0.0077|
 |boolq        |      1|acc     |0.8599|_  |0.0061|
+|hellaswag    |      0|acc     |0.6520|_  |0.0048|
+|             |       |acc_norm|0.8357|_  |0.0037|
+|openbookqa   |      0|acc     |0.3440|_  |0.0213|
+|             |       |acc_norm|0.4580|_  |0.0223|
+|piqa         |      0|acc     |0.8199|_  |0.0090|
+|             |       |acc_norm|0.8319|_  |0.0087|
+|winogrande   |      0|acc     |0.7482|_  |0.0122|
 ```
+Average: 0.7422
 ## TruthfulQA
 ```
 |    Task     |Version|Metric|Value |   |Stderr|
 |-------------|------:|------|-----:|---|-----:|
+|truthfulqa_mc|      1|mc1   |0.3941|_  |0.0171|
+|             |       |mc2   |0.5698|_  |0.0154|
 ```
 <!-- original-model-card end -->