This is a sentence-transformers model finetuned from sentence-transformers/all-MiniLM-L6-v2 on the monster-sanctuary-conversations dataset. It maps sentences & paragraphs to a 384-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

Model Details

Model Description

Model Type: Sentence Transformer
Base model: sentence-transformers/all-MiniLM-L6-v2
Maximum Sequence Length: 256 tokens
Output Dimensionality: 384 dimensions
Similarity Function: Cosine Similarity
Training Dataset:
- monster-sanctuary-conversations
Language: en
License: apache-2.0

Model Sources

Documentation: Sentence Transformers Documentation
Repository: Sentence Transformers on GitHub
Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 256, 'do_lower_case': False, 'architecture': 'BertModel'})
  (1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
  (2): Normalize()
)

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load this model and run inference.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("woodman231/minilm-l6-monster-sanctuary")
# Run inference
sentences = [
    'Which monster can be hatched from the item Mega Potion?',
    'The item Mega Potion does not hatch into any monster.',
    'The cost of Ocarina+3 is 3500 gold.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[ 1.0000,  0.9527, -0.0061],
#         [ 0.9527,  1.0000, -0.0173],
#         [-0.0061, -0.0173,  1.0000]])

Training Details

Training Dataset

monster-sanctuary-conversations

Dataset: monster-sanctuary-conversations at fb621dc
Size: 9,103 training samples
Columns: anchor and positive
Approximate statistics based on the first 1000 samples:
anchor positive
type string string
details
min: 8 tokens
mean: 12.92 tokens
max: 18 tokens
min: 9 tokens
mean: 26.31 tokens
max: 256 tokens

Samples:

anchor	positive
`Which monster can be hatched from the item Fan+3?`	`The item Fan+3 does not hatch into any monster.`
`Which items can Infinity Flame+2 be upgraded to?`	`The item Infinity Flame+2 can be upgraded to: Infinity Flame+3. It requires the following materials: 1 x Crimson Gem, 1 x Azure Gem, 1 x Verdant Gem.`
`Which monsters drop the item Lucky Clover?`	`The item Lucky Clover is not dropped by any monsters.`

Loss: MultipleNegativesRankingLoss with these parameters:
json
1{ 2 "scale": 20.0, 3 "similarity_fct": "cos_sim", 4 "gather_across_devices": false 5}

Evaluation Dataset

monster-sanctuary-conversations

Dataset: monster-sanctuary-conversations at fb621dc
Size: 1,012 evaluation samples
Columns: anchor and positive
Approximate statistics based on the first 1000 samples:
anchor positive
type string string
details
min: 8 tokens
mean: 12.95 tokens
max: 18 tokens
min: 10 tokens
mean: 29.16 tokens
max: 256 tokens

Samples:

anchor	positive
`What Magical Damage Elements does Toxiquus deal?`	`The magical damage elements dealt by Toxiquus are:[object Object]- Wind[object Object]- Earth.`
`What is the cost of the item Cestus+2?`	`The cost of Cestus+2 is 850 gold.`
`What item can Steel be used to upgrade?`	The item Steel can be used to upgrade the following items:[object Object]- Abyssal Sword+3 (requires 1)[object Object]- Abyssal Sword+4 (requires 3)[object Object]- Belt+3 (requires 2)[object Object]- Belt+4 (requires 2)[object Object]- Bow+3 (requires 2)[object Object]- Bracelet+3 (requires 2)[object Object]- Bracelet+4 (requires 4)[object Object]- Bracer+3 (requires 2)[object Object]- Bracer+4 (requires 4)[object Object]- Buckler+3 (requires 2)[object Object]- Cauldron+3 (requires 2)[object Object]- Cestus+3 (requires 2)[object Object]- Cestus+4 (requires 4)[object Object]- Charging Sphere+3 (requires 4)[object Object]- Claws+3 (requires 2)[object Object]- Claws+4 (requires 4)[object Object]- Dumbbell+3 (requires 2)[object Object]- Dumbbell+4 (requires 2)[object Object]- Fang+3 (requires 1)[object Object]- Fang+4 (requires 2)[object Object]- Gauntlet+3 (requires 2)[object Object]- Gauntlet+4 (requires 4)[object Object]- Hammer+3 (requires 2)[object Object]- Hammer+4 (requires 4)[object Object]- Heavy Mace+3 (requires 2)[object Object]- Heavy Mace+4 (requires 4)[object Object]- Helmet+3 (requires 2)[object Object]- Helmet+4 (requires 3)[object Object]- Hourglass+3 (requires 1)[object Object]- Hourglass+4 (requires 2)[object Object]- Katana+3 (requires 1)[object Object]- Katana+4 (requires 2)[object Object]- Katar+3 (requires 2)[object Object]- Katar+4 (requires 4)[object Object]- Kunai+3 (requires 2)[object Object]- Kunai+4 (requires 4)[object Object]- Large Shield+3 (requires 1)[object Object]- Large Shie...

Loss: MultipleNegativesRankingLoss with these parameters:
json
1{ 2 "scale": 20.0, 3 "similarity_fct": "cos_sim", 4 "gather_across_devices": false 5}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: epoch
per_device_train_batch_size: 32
per_device_eval_batch_size: 32
learning_rate: 2e-05
num_train_epochs: 5
warmup_ratio: 0.1
fp16: True
batch_sampler: no_duplicates

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: epoch
prediction_loss_only: True
per_device_train_batch_size: 32
per_device_eval_batch_size: 32
per_gpu_train_batch_size: None
per_gpu_eval_batch_size: None
gradient_accumulation_steps: 1
eval_accumulation_steps: None
torch_empty_cache_steps: None
learning_rate: 2e-05
weight_decay: 0.0
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1e-08
max_grad_norm: 1.0
num_train_epochs: 5
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: {}
warmup_ratio: 0.1
warmup_steps: 0
log_level: passive
log_level_replica: warning
log_on_each_node: True
logging_nan_inf_filter: True
save_safetensors: True
save_on_each_node: False
save_only_model: False
restore_callback_states_from_checkpoint: False
no_cuda: False
use_cpu: False
use_mps_device: False
seed: 42
data_seed: None
jit_mode_eval: False
bf16: False
fp16: True
fp16_opt_level: O1
half_precision_backend: auto
bf16_full_eval: False
fp16_full_eval: False
tf32: None
local_rank: 0
ddp_backend: None
tpu_num_cores: None
tpu_metrics_debug: False
debug: []
dataloader_drop_last: False
dataloader_num_workers: 0
dataloader_prefetch_factor: None
past_index: -1
disable_tqdm: False
remove_unused_columns: True
label_names: None
load_best_model_at_end: False
ignore_data_skip: False
fsdp: []
fsdp_min_num_params: 0
fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
fsdp_transformer_layer_cls_to_wrap: None
accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
parallelism_config: None
deepspeed: None
label_smoothing_factor: 0.0
optim: adamw_torch_fused
optim_args: None
adafactor: False
group_by_length: False
length_column_name: length
project: huggingface
trackio_space_id: trackio
ddp_find_unused_parameters: None
ddp_bucket_cap_mb: None
ddp_broadcast_buffers: False
dataloader_pin_memory: True
dataloader_persistent_workers: False
skip_memory_metrics: True
use_legacy_prediction_loop: False
push_to_hub: False
resume_from_checkpoint: None
hub_model_id: None
hub_strategy: every_save
hub_private_repo: None
hub_always_push: False
hub_revision: None
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
include_inputs_for_metrics: False
include_for_metrics: []
eval_do_concat_batches: True
fp16_backend: auto
push_to_hub_model_id: None
push_to_hub_organization: None
mp_parameters:
auto_find_batch_size: False
full_determinism: False
torchdynamo: None
ray_scope: last
ddp_timeout: 1800
torch_compile: False
torch_compile_backend: None
torch_compile_mode: None
include_tokens_per_second: False
include_num_input_tokens_seen: no
neftune_noise_alpha: None
optim_target_modules: None
batch_eval_metrics: False
eval_on_start: False
use_liger_kernel: False
liger_kernel_config: None
eval_use_gather_object: False
average_tokens_across_devices: True
prompts: None
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	Validation Loss
0.1754	50	0.0223	-
0.3509	100	0.007	-
0.5263	150	0.0037	-
0.7018	200	0.0028	-
0.8772	250	0.0024	-
1.0	285	-	0.0012
1.0526	300	0.0029	-
1.2281	350	0.0011	-
1.4035	400	0.0018	-
1.5789	450	0.0008	-
1.7544	500	0.0013	-
1.9298	550	0.0007	-
2.0	570	-	0.0007
2.1053	600	0.0011	-
2.2807	650	0.0011	-
2.4561	700	0.0011	-
2.6316	750	0.0009	-
2.8070	800	0.0008	-
2.9825	850	0.0008	-
3.0	855	-	0.0004
3.1579	900	0.0009	-
3.3333	950	0.0017	-
3.5088	1000	0.0014	-
3.6842	1050	0.0008	-
3.8596	1100	0.0007	-
4.0	1140	-	0.0004
4.0351	1150	0.002	-
4.2105	1200	0.0008	-
4.3860	1250	0.0009	-
4.5614	1300	0.0011	-
4.7368	1350	0.0005	-
4.9123	1400	0.0008	-
5.0	1425	-	0.0004

Framework Versions

Python: 3.10.12
Sentence Transformers: 5.1.2
Transformers: 4.57.1
PyTorch: 2.9.1+cu128
Accelerate: 1.11.0
Datasets: 4.4.1
Tokenizers: 0.22.1

Citation

BibTeX

Sentence Transformers

bibtex

@inproceedings{reimers-2019-sentence-bert,
    title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
    author = "Reimers, Nils and Gurevych, Iryna",
    booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
    month = "11",
    year = "2019",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/1908.10084",
}

MultipleNegativesRankingLoss

bibtex

@misc{henderson2017efficient,
    title={Efficient Natural Language Response Suggestion for Smart Reply},
    author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
    year={2017},
    eprint={1705.00652},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}

	anchor	positive
type	string	string
details	min: 8 tokens mean: 12.92 tokens max: 18 tokens	min: 9 tokens mean: 26.31 tokens max: 256 tokens