Create README.md

README.md

@@ -0,0 +1,86 @@
+---
+license: cc-by-nc-4.0
+language:
+- en
+pipeline_tag: text-generation
+base_model: Sao10K/L3-8B-Stheno-v3.3-32K
+---
+
+# QuantFactory/L3-8B-Stheno-v3.3-32K-GGUF
+This is quantized version of [Sao10K/L3-8B-Stheno-v3.3-32K](https://huggingface.co/Sao10K/L3-8B-Stheno-v3.3-32K) created using llama.cpp
+
+# Model Description
+
+Trained with compute from [Backyard.ai](https://backyard.ai/)
+
+Training Details:
+<br>Trained at 8K Context -> Expanded to 32K Context with PoSE training.
+
+Dataset Modifications:
+<br>\- Further Cleaned up Roleplaying Samples -> Quality Check
+<br>\- Removed Low Quality Samples from Manual Check -> Increased Baseline Quality Floor
+<br>\- More Creative Writing Samples -> 2x Samples
+<br>\- Remade and Refined Detailed Instruct Data
+
+Notes:
+<br>\- Training run is much less aggressive than previous Stheno versions.
+<br>\- This model works when tested in bf16 with the same configs as within the file.
+<br>\- I do not know the effects quantisation has on it.
+<br>\- Roleplays pretty well. Feels nice in my opinion.
+<br>\- It has some issues on long context understanding and reasoning. Much better vs rope scaling normally though, so that is a plus.
+<br>\- Reminder, this isn't a native 32K model. It has it's issues, but it's coherent and working well.
+
+Sanity Check // Needle in a Haystack Results:
+<br>\- This is not as complex as RULER or NIAN, but it's a basic evaluator. Some improper train examples had Haystack scores ranging from Red to Orange for most of the extended contexts.
+![Results](https://huggingface.co/Sao10K/L3-8B-Stheno-v3.3-32K/resolve/main/haystack.png)
+
+Wandb Run:
+![Wandb](https://huggingface.co/Sao10K/L3-8B-Stheno-v3.3-32K/resolve/main/wandb.png)
+
+---
+
+Relevant Axolotl Configurations:
+<br>-> Taken from [winglian/Llama-3-8b-64k-PoSE](https://huggingface.co/winglian/Llama-3-8b-64k-PoSE)
+<br>\- I tried to find my own configs, hours of tinkering but the one he used worked best, so I stuck to it.
+<br>\- 2M Rope Theta had the best loss results during training compared to other values.
+<br>\- Leaving it at 500K rope wasn't that much worse, but 4M and 8M Theta made the grad_norm values worsen even if loss drops fast.
+<br>\- Mixing in Pretraining Data was a PITA. Made it a lot worse with formatting.
+<br>\- Pretraining / Noise made it worse at Haystack too? It wasn't all Green, Mainly Oranges.
+<br>\- Improper / Bad Rope Theta shows in Grad_Norm exploding to thousands. It'll drop to low values alright, but it's a scary fast drop even with gradient clipping.
+
+```
+sequence_len: 8192
+use_pose: true
+pose_max_context_len: 32768
+
+overrides_of_model_config:
+  rope_theta: 2000000.0
+  max_position_embeddings: 32768
+
+  # peft_use_dora: true
+adapter: lora
+peft_use_rslora: true
+lora_model_dir:
+lora_r: 256
+lora_alpha: 256
+lora_dropout: 0.1
+lora_target_linear: true
+lora_target_modules:
+  - gate_proj
+  - down_proj
+  - up_proj
+  - q_proj
+  - v_proj
+  - k_proj
+  - o_proj
+
+warmup_steps: 80
+gradient_accumulation_steps: 6
+micro_batch_size: 1
+num_epochs: 2
+optimizer: adamw_bnb_8bit
+lr_scheduler: cosine_with_min_lr
+learning_rate: 0.00004
+lr_scheduler_kwargs:
+    min_lr: 0.000004
+```