End of training
Browse files- README.md +10 -2
- config.json +38 -0
README.md
CHANGED
|
@@ -2,11 +2,19 @@
|
|
| 2 |
license: apache-2.0
|
| 3 |
library_name: peft
|
| 4 |
tags:
|
|
|
|
|
|
|
| 5 |
- trl
|
| 6 |
- sft
|
| 7 |
- generated_from_trainer
|
| 8 |
datasets:
|
| 9 |
-
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
base_model: jan-hq/stealth-v1.3
|
| 11 |
model-index:
|
| 12 |
- name: stealth-rag-v1
|
|
@@ -18,7 +26,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 18 |
|
| 19 |
# stealth-rag-v1
|
| 20 |
|
| 21 |
-
This model is a fine-tuned version of [jan-hq/stealth-v1.3](https://huggingface.co/jan-hq/stealth-v1.3) on the
|
| 22 |
It achieves the following results on the evaluation set:
|
| 23 |
- Loss: 1.3883
|
| 24 |
|
|
|
|
| 2 |
license: apache-2.0
|
| 3 |
library_name: peft
|
| 4 |
tags:
|
| 5 |
+
- alignment-handbook
|
| 6 |
+
- generated_from_trainer
|
| 7 |
- trl
|
| 8 |
- sft
|
| 9 |
- generated_from_trainer
|
| 10 |
datasets:
|
| 11 |
+
- jan-hq/rag_dataset_1200_binarized
|
| 12 |
+
- jan-hq/rag_dataset_12000_binarized
|
| 13 |
+
- jan-hq/rag_hallucination_dataset_1000_binarized
|
| 14 |
+
- jan-hq/rag_full_20000_binarized
|
| 15 |
+
- jan-hq/bagel_sft_binarized
|
| 16 |
+
- jan-hq/dolphin_binarized
|
| 17 |
+
- jan-hq/openhermes_binarized
|
| 18 |
base_model: jan-hq/stealth-v1.3
|
| 19 |
model-index:
|
| 20 |
- name: stealth-rag-v1
|
|
|
|
| 26 |
|
| 27 |
# stealth-rag-v1
|
| 28 |
|
| 29 |
+
This model is a fine-tuned version of [jan-hq/stealth-v1.3](https://huggingface.co/jan-hq/stealth-v1.3) on the jan-hq/rag_dataset_1200_binarized, the jan-hq/rag_dataset_12000_binarized, the jan-hq/rag_hallucination_dataset_1000_binarized, the jan-hq/rag_full_20000_binarized, the jan-hq/bagel_sft_binarized, the jan-hq/dolphin_binarized and the jan-hq/openhermes_binarized datasets.
|
| 30 |
It achieves the following results on the evaluation set:
|
| 31 |
- Loss: 1.3883
|
| 32 |
|
config.json
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "jan-hq/stealth-v1.3",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"MistralForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": 1,
|
| 8 |
+
"eos_token_id": 2,
|
| 9 |
+
"hidden_act": "silu",
|
| 10 |
+
"hidden_size": 4096,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"intermediate_size": 14336,
|
| 13 |
+
"max_position_embeddings": 32768,
|
| 14 |
+
"model_type": "mistral",
|
| 15 |
+
"num_attention_heads": 32,
|
| 16 |
+
"num_hidden_layers": 32,
|
| 17 |
+
"num_key_value_heads": 8,
|
| 18 |
+
"quantization_config": {
|
| 19 |
+
"bnb_4bit_compute_dtype": "float16",
|
| 20 |
+
"bnb_4bit_quant_type": "nf4",
|
| 21 |
+
"bnb_4bit_use_double_quant": false,
|
| 22 |
+
"llm_int8_enable_fp32_cpu_offload": false,
|
| 23 |
+
"llm_int8_has_fp16_weight": false,
|
| 24 |
+
"llm_int8_skip_modules": null,
|
| 25 |
+
"llm_int8_threshold": 6.0,
|
| 26 |
+
"load_in_4bit": true,
|
| 27 |
+
"load_in_8bit": false,
|
| 28 |
+
"quant_method": "bitsandbytes"
|
| 29 |
+
},
|
| 30 |
+
"rms_norm_eps": 1e-05,
|
| 31 |
+
"rope_theta": 10000.0,
|
| 32 |
+
"sliding_window": 4096,
|
| 33 |
+
"tie_word_embeddings": false,
|
| 34 |
+
"torch_dtype": "bfloat16",
|
| 35 |
+
"transformers_version": "4.36.2",
|
| 36 |
+
"use_cache": true,
|
| 37 |
+
"vocab_size": 32000
|
| 38 |
+
}
|