Hugging Face's logo

keyvan-ai
/
Mankei-326M-Reranker

Text Classification
Transformers
Safetensors
German
llama
feature-extraction
reranker
cross-encoder
german
retrieval
rag
on-prem
text-embeddings-inference
Model card Files
xet
 Community
Keyven commited on
Commit
3a89c44
·
verified ·
1 Parent(s): 444a952

Reranker v3: 326M, listwise (8 Kandidaten, Hard-Negatives) - 92,8% Acc@1, schlaegt bge-reranker-v2-m3

Browse files
Files changed (4) hide show
  1. config.json +8 -22
  2. head.pt +2 -2
  3. model.safetensors +2 -2
  4. tokenizer.json +14 -2
config.json CHANGED
@@ -1,46 +1,32 @@
1
  {
2
- "add_cross_attention": false,
3
  "architectures": [
4
  "LlamaModel"
5
  ],
6
  "attention_bias": false,
7
  "attention_dropout": 0.0,
8
  "bos_token_id": 0,
9
- "cross_attention_hidden_size": null,
10
- "decoder_start_token_id": null,
11
  "dtype": "bfloat16",
12
  "eos_token_id": 1,
13
- "finetuning_task": null,
14
  "head_dim": 64,
15
  "hidden_act": "silu",
16
- "hidden_size": 1536,
17
  "initializer_range": 0.02,
18
- "intermediate_size": 5248,
19
- "is_decoder": false,
20
- "max_position_embeddings": 4096,
21
  "mlp_bias": false,
22
  "model_type": "llama",
23
- "num_attention_heads": 24,
24
- "num_hidden_layers": 32,
25
- "num_key_value_heads": 6,
26
  "pad_token_id": 1,
27
- "prefix": null,
28
  "pretraining_tp": 1,
29
- "pruned_heads": {},
30
  "rms_norm_eps": 1e-05,
31
  "rope_parameters": {
32
- "rope_theta": 500000.0,
33
  "rope_type": "default"
34
  },
35
- "sep_token_id": null,
36
- "task_specific_params": null,
37
- "tf_legacy_loss": false,
38
- "tie_encoder_decoder": false,
39
  "tie_word_embeddings": true,
40
- "tokenizer_class": null,
41
- "torchscript": false,
42
  "transformers_version": "5.15.0",
43
- "use_bfloat16": false,
44
- "use_cache": true,
45
  "vocab_size": 32768
46
  }
 
1
  {
 
2
  "architectures": [
3
  "LlamaModel"
4
  ],
5
  "attention_bias": false,
6
  "attention_dropout": 0.0,
7
  "bos_token_id": 0,
 
 
8
  "dtype": "bfloat16",
9
  "eos_token_id": 1,
 
10
  "head_dim": 64,
11
  "hidden_act": "silu",
12
+ "hidden_size": 960,
13
  "initializer_range": 0.02,
14
+ "intermediate_size": 2560,
15
+ "max_position_embeddings": 2048,
 
16
  "mlp_bias": false,
17
  "model_type": "llama",
18
+ "num_attention_heads": 15,
19
+ "num_hidden_layers": 30,
20
+ "num_key_value_heads": 5,
21
  "pad_token_id": 1,
 
22
  "pretraining_tp": 1,
 
23
  "rms_norm_eps": 1e-05,
24
  "rope_parameters": {
25
+ "rope_theta": 100000.0,
26
  "rope_type": "default"
27
  },
 
 
 
 
28
  "tie_word_embeddings": true,
 
 
29
  "transformers_version": "5.15.0",
30
+ "use_cache": false,
 
31
  "vocab_size": 32768
32
  }
head.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2edfc785218fb94de9cfea84c90b62cc9f58469cebdc6d84b214d8d2dcc53a18
3
- size 7949
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ce0cf93891cbee260a7acfb9c5d66bf3d9c13c68e81fd9eb5cac99838b543edf
3
+ size 5645
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:24386449555b5fc07c9819cb000da2dae9d1614e68d31b232d0196369b0e0b08
3
- size 2026080120
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:34566de5240d16995f2830c00ea5981d5f3013f165b2c75a440100b7c49375d8
3
+ size 652884656
tokenizer.json CHANGED
@@ -1,7 +1,19 @@
1
  {
2
  "version": "1.0",
3
- "truncation": null,
4
- "padding": null,
 
 
 
 
 
 
 
 
 
 
 
 
5
  "added_tokens": [
6
  {
7
  "id": 0,
 
1
  {
2
  "version": "1.0",
3
+ "truncation": {
4
+ "direction": "Right",
5
+ "max_length": 192,
6
+ "strategy": "LongestFirst",
7
+ "stride": 0
8
+ },
9
+ "padding": {
10
+ "strategy": "BatchLongest",
11
+ "direction": "Right",
12
+ "pad_to_multiple_of": null,
13
+ "pad_id": 1,
14
+ "pad_type_id": 0,
15
+ "pad_token": "<|end_of_text|>"
16
+ },
17
  "added_tokens": [
18
  {
19
  "id": 0,