Commit
•
e4a68c6
1
Parent(s):
eddfbc0
Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
datasets:
|
| 3 |
+
- ehartford/WizardLM_evol_instruct_V2_196k_unfiltered_merged_split
|
| 4 |
+
- jondurbin/airoboros-gpt4-1.4.1
|
| 5 |
+
- openai/summarize_from_feedback
|
| 6 |
+
- ehartford/wizard_vicuna_70k_unfiltered
|
| 7 |
+
language:
|
| 8 |
+
- en
|
| 9 |
+
tags:
|
| 10 |
+
- llama
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
Trained on a flavorful melange of the WizardLM, Airoboros, and Wizard Vicuna datasets.
|
| 14 |
+
This model was trained using both linear and NTK-aware RoPE scaling in tandem. When loading, ensure that `compress_pos_emb` (or `scale`) is set to 2, and `alpha_value` is set to 4. *Both* values must be set.
|
| 15 |
+
|
| 16 |
+
Expect context length of up to 8192 to work for sure. It will probably maintain coherence into the ~12k range, but I have not tested that.
|