Add 8b version of context relevance
Browse files
context_relevance/lora/granite-3.3-8b-instruct/adapter_config.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "ibm-granite/granite-3.3-8b-instruct",
|
| 5 |
+
"bias": "none",
|
| 6 |
+
"eva_config": null,
|
| 7 |
+
"exclude_modules": null,
|
| 8 |
+
"fan_in_fan_out": false,
|
| 9 |
+
"inference_mode": true,
|
| 10 |
+
"init_lora_weights": true,
|
| 11 |
+
"layer_replication": null,
|
| 12 |
+
"layers_pattern": null,
|
| 13 |
+
"layers_to_transform": null,
|
| 14 |
+
"loftq_config": {},
|
| 15 |
+
"lora_alpha": 32,
|
| 16 |
+
"lora_bias": false,
|
| 17 |
+
"lora_dropout": 0.1,
|
| 18 |
+
"megatron_config": null,
|
| 19 |
+
"megatron_core": "megatron.core",
|
| 20 |
+
"modules_to_save": null,
|
| 21 |
+
"peft_type": "LORA",
|
| 22 |
+
"r": 32,
|
| 23 |
+
"rank_pattern": {},
|
| 24 |
+
"revision": null,
|
| 25 |
+
"target_modules": [
|
| 26 |
+
"q_proj",
|
| 27 |
+
"k_proj",
|
| 28 |
+
"v_proj"
|
| 29 |
+
],
|
| 30 |
+
"task_type": "CAUSAL_LM",
|
| 31 |
+
"use_dora": false,
|
| 32 |
+
"use_rslora": false
|
| 33 |
+
}
|
context_relevance/lora/granite-3.3-8b-instruct/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:56311cd911b68a1972f238f50f4671dbbb35335da0a2ec29fecc2e843a17fa81
|
| 3 |
+
size 94404160
|
context_relevance/lora/granite-3.3-8b-instruct/io.yaml
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Model name string, or null to use whatever is provided in the chat completion request
|
| 2 |
+
model: ~
|
| 3 |
+
# JSON schema of the model's output
|
| 4 |
+
response_format: |
|
| 5 |
+
{
|
| 6 |
+
"title": "ContextRelevanceOutput",
|
| 7 |
+
"type": "object",
|
| 8 |
+
"properties": {
|
| 9 |
+
"context_relevance": {
|
| 10 |
+
"type": "string",
|
| 11 |
+
"description": "Context relevancy judgment.",
|
| 12 |
+
"enum": ["relevant", "irrelevant", "partially relevant"]
|
| 13 |
+
}
|
| 14 |
+
},
|
| 15 |
+
"required": ["context_relevance"]
|
| 16 |
+
}
|
| 17 |
+
# Additional turn of instructions to add to the chat
|
| 18 |
+
instruction: |
|
| 19 |
+
DOCUMENT: {document_content}
|
| 20 |
+
# Data transformations to perform during post-processing
|
| 21 |
+
transformations:
|
| 22 |
+
- type: likelihood
|
| 23 |
+
categories_to_values:
|
| 24 |
+
"relevant": 1.0
|
| 25 |
+
"irrelevant": 0.0
|
| 26 |
+
"partially relevant": 0.5
|
| 27 |
+
input_path: ["context_relevance"]
|
| 28 |
+
output_name: ~ # Null == no change
|
| 29 |
+
parameters: ~
|
| 30 |
+
sentence_boundaries: false
|