Upload folder using huggingface_hub
Browse files- README.md +37 -0
- config.json +33 -0
README.md
ADDED
|
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- moe
|
| 5 |
+
- mergekit
|
| 6 |
+
- vortexmergekit
|
| 7 |
+
- OEvortex/HelpingAI-HelpingAI-unvelite
|
| 8 |
+
- OEvortex/lite-hermes
|
| 9 |
+
- OEvortex/HelpingAI-Lite
|
| 10 |
+
- OEvortex/HelpingAI-Lite-1.5T
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# HelpingAI-Lite-4x1b
|
| 14 |
+
|
| 15 |
+
Hey there! 👋 Welcome to the HelpingAI-Lite-4x1b! This is a MOE brought together using the awesome [VortexMerge kit](https://colab.research.google.com/drive/1YjcvCLuNG1PK7Le6_4xhVU5VpzTwvGhk#scrollTo=UG5H2TK4gVyl).
|
| 16 |
+
|
| 17 |
+
Let's see what we've got in this model:
|
| 18 |
+
* [OEvortex/HelpingAI-HelpingAI-unvelite](https://huggingface.co/OEvortex/HelpingAI-HelpingAI-unvelite) 🚀
|
| 19 |
+
* [OEvortex/lite-hermes](https://huggingface.co/OEvortex/lite-hermes) 🚀
|
| 20 |
+
* [OEvortex/HelpingAI-Lite](https://huggingface.co/OEvortex/HelpingAI-Lite) 🚀
|
| 21 |
+
* [OEvortex/HelpingAI-Lite-1.5T](https://huggingface.co/OEvortex/HelpingAI-Lite-1.5T) 🚀
|
| 22 |
+
|
| 23 |
+
## 🧩 Configuration
|
| 24 |
+
|
| 25 |
+
```yaml
|
| 26 |
+
base_model: OEvortex/HelpingAI-Lite
|
| 27 |
+
dtype: float16
|
| 28 |
+
gate_mode: cheap_embed
|
| 29 |
+
experts:
|
| 30 |
+
- source_model: OEvortex/HelpingAI-HelpingAI-unvelite
|
| 31 |
+
positive_prompts: ["You are an helpful unbounded, uncensored assistant."]
|
| 32 |
+
- source_model: OEvortex/lite-hermes
|
| 33 |
+
positive_prompts: ["You are helpful assistant."]
|
| 34 |
+
- source_model: OEvortex/HelpingAI-Lite
|
| 35 |
+
positive_prompts: ["You are a coding assistant."]
|
| 36 |
+
- source_model: OEvortex/HelpingAI-Lite-1.5T
|
| 37 |
+
positive_prompts: ["You are helpful and general-purpose assistant."]
|
config.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "OEvortex/HelpingAI-Lite",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"MixtralForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"bos_token_id": 1,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "silu",
|
| 11 |
+
"hidden_size": 2048,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 5632,
|
| 14 |
+
"max_position_embeddings": 2048,
|
| 15 |
+
"model_type": "mixtral",
|
| 16 |
+
"num_attention_heads": 32,
|
| 17 |
+
"num_experts_per_tok": 2,
|
| 18 |
+
"num_hidden_layers": 22,
|
| 19 |
+
"num_key_value_heads": 4,
|
| 20 |
+
"num_local_experts": 4,
|
| 21 |
+
"output_router_logits": false,
|
| 22 |
+
"pretraining_tp": 1,
|
| 23 |
+
"rms_norm_eps": 1e-05,
|
| 24 |
+
"rope_scaling": null,
|
| 25 |
+
"rope_theta": 10000.0,
|
| 26 |
+
"router_aux_loss_coef": 0.001,
|
| 27 |
+
"sliding_window": null,
|
| 28 |
+
"tie_word_embeddings": false,
|
| 29 |
+
"torch_dtype": "float16",
|
| 30 |
+
"transformers_version": "4.39.0.dev0",
|
| 31 |
+
"use_cache": true,
|
| 32 |
+
"vocab_size": 32000
|
| 33 |
+
}
|