--- tags: - sentence-transformers - sentence-similarity - feature-extraction - generated_from_trainer - dataset_size:200 - loss:MatryoshkaLoss - loss:MultipleNegativesRankingLoss base_model: Snowflake/snowflake-arctic-embed-m widget: - source_sentence: How many times should you squeeze the booty while lifting the leg? sentences: - 'the elbow then we''re gonna punch it forward bend it and extend behind you oh good you guys Bend use that core extend forward and return and Ben big exhale forward and back check those hips that they''re equal you''re not sinking to one side three more use that belly you guys if you''re fill in the back just place that knee down two all right hold it up there on one use the belly use the core hold hold hold drop the leg and lift just five squeeze the booty four three two and one and bring it in give yourself a little round through your back all right we''re staying in this kneeling position and you''re gonna come up you''re on the lighter spring in this unstable position so be careful you''re gonna sweep this arm forward' - 'you guys can exhale press curl up pause inhale let''s reach the right leg out to a challenging level exhale the pull back in inhale left leg extends exhale in inhale out and exhale back think engaging our obliques to pull each leg back in inhale and exhale let''s do one more right one more left arms can go up heads can go down relax a feet give yourselves a stretch like send her arms back up to the ceiling you guys can bring your knees up into your tabletop position again again one abdominal curl let''s exhale press curl up pause at the top inhale reach full flakes out to a challenging level exhale scoop into your low ABS to pull in inhale out and exhaling as you pull back in inhale away exhale to pull think draw your abs into' - 'into that abdominal curl inhale lower your right leg straight down towards your foot bar exhale lift left leg lowers and exhale lift as always think single legs engaging your obliques to pull each leg back up to the ceiling we''ll do three more one more right one more left bend your knees arms up head down relax the feet feel free to give yourselves a little stretch last thing will be that double leg lowering so let''s extend our arms up to the ceiling knees can go up into our table top position legs go straight up to the ceiling exhale press curl yourselves back up pause inhale lower both legs down to a challenging level and then exhale scoop into your low ABS to lift inhale lower and exhaling as you lift think drawing' - source_sentence: What position should the feet be in during the exercise described in the context? sentences: - 'by our sides hold it down there tricep press keeping the elbows tight and reach just two more this way then we''re going to be doing like an L with our tricep so open your right tricep out to the side left tricep towards the ceiling and then come back to center now open the left tricep out to the side right tricep up towards the ceiling and down good meanwhile you''re keeping your shoulders nice and stable inhale and exhale through it inhale to bend exhale to straighten if you feel that low back Bend those knees in closer last set of each side one squeeze one squeeze find your Center and release good open those arms out to the side and rock your knees over to one side for a little recovery big deep breath' - 'equal 3 2 hold it on one drop the hips an inch and up down and up for eight s six squeeze squeeze squeeze keep the core tight lower your hips a little if you''re feeling the low back 4 3 2 one place the foot down both hips lift nice and high inhale exhale rolling down the spine and and finishing in your neutral windshield wiper the legs good job you guys back of the legs should feel nice and warm we''re going to stay on our heels hip distance apart headrest can pop up now for Comfort if you''d like we want to be in a neutral spine space in the low back inhale to prepare exhale press out on your heels and back in tailbone should be nice and heavy you can place your hands on your hips to check you''re neutral after we''ve' - 'come halfway in pulse it down nice wide pull squeezing the heels 8 7 6 5 4 3 2 One reach it out to finish and come back in woo shake out those legs they should be on fire all right we''re coming all the way up to a seated position we''re going to switch to just a medium tension spring so that''s going to be a Blue Spring on this machine you can also do a light spring as well so that would be a blue and white if you''re wanting a little more weight we''re working our arms so either medium or medium and a light you''re going to go kneeling or seated I''m going to take a kneeling position and grab a hold of my straps if you''re seated you can sit in a crisscross position kneeling make sure that you''re aware of your body Before' - source_sentence: How should you position your left foot and right knee during the exercise? sentences: - 'everything up together as one unit as you bring the carriage back in again inhale press out hips drop down to a hover you still have that natural curve of the low back and then you''re going to squeeze and lift up keeping that neutral spine the whole time inhale out exhale lift if you feel the low back don''t take your hips quite as high if this arm reach is too much and you want that assistance just take your arms back down to the carriage two more exhale lift keep that rib cage zipped use that cord to help you last one lift hold it in we''re doing just ten little pulses in and in lower those hips if you need to for the low back we want to just focus on those hamstrings and glutes working four three two' - 'side circles here for four three two one bend your knee relax your leg take that strap off your foot and then once you''re gonna take that strap off keep it into your hand we''re going to switch now so the strap is in your left hand we''re going to rotate to the right side I''m going to kneel like Center my reformer my left foot is going to go into my headrest and then my right knee down again if this is too heavy for you guys feel free to switch to a Blue Spring a medium spring instead of this red arms are going to open up wide to the sides long Loop is going to be lighter for your arm short Loop is heavier slight little bench your elbows on an exhale arms go up overhead inhale return down exhale fingertips together' - 'Carriage I''m going to take my left foot down in between both hands and then I''m going to kneel nice and Tall both elbows are going to be bent Palms are flipping up and we''re going to Exhale just reach your right arm out inhale back exhale out inhale back making sure that that shoulder stays back as the arm reaches away three more two and then last one relax that arm you can rotate back hang that strap up we''re going to step off and then we are going to do the same thing on the other side so kneel on your reformers face your straps you guys are going to grab the right strap into your hand this time long Loop or short loop again up to you I have my long Loop hand is on the outside frame of the carriage Palm is' - source_sentence: How does the movement sequence aim to engage different muscle groups? sentences: - 'and then place it back down so we''re gonna do one runner in shoot it out hold it there and then jump off Pike and place it back down so again run push it out jump off the right leg Pike place it down good run in exhale squeeze Pike it up and down and this is where we get that heart rate up and that quad really burning good so full body movement here shoulders chest are supporting big time in that Pike core is working at the peak of that Pike good Pike and lift four more in squeeze Pike and lift and three Pike and lit and two [Music] yes and a one good we are going to step your right foot in front of the left so you can take a look if you''d like cross it all the way over in front to the left we want to stay bent through your right' - 'are going to start first by straddling your reformers your feet are on the floor we''re going to take our hands to our shoulder blocks we''re just going to do a quick stretch before we get moving so I''m going to bend my knees slightly I''m going to inhale press my Carriage out let my chest drop down in between my arms and then my exhale I''m going to tuck my pelvis around through my spine to come back in inhale press out let your chest drop down X tail tuck around to come in we have two more again after this we''re going to get a really good workout in today last one and then round and come in all right now once we bring it back in we''re going to take our knees onto our carriages and then our hands are going to go into our' - '[Music] hey guys welcome back I''m Dez and today I''m taking you through a 45 minute full body Pilates reformer workout this workout does include some exercises using the platform extender however it is not required as you could stand on your wood platform or something similar depending on your reformer model either way I hope you''ll join me for this fun and challenging workout and let''s get started okay you guys we''re starting on three heavy Springs today or whatever you prefer for footwork and we''re going to be laying down on our backs I want your headress to be down flat to start today we''re going to start with your heels on the bar neutral spine taking an inhale right here to prepare I want you to just' - source_sentence: What muscle groups are primarily engaged during the lunge exercise described in the context? sentences: - 'hips and we''re gonna lunge it down the weight is going to feel a little bit light we''re working more stabilizers to start here stabilizers through the hips ankles knees all the things okay lunge it down I like to put my foot right up against that edge to help me have a nice grip toes are off of the carriage and then coming up to squeeze up on a straight standing leg squeeze up through the glute down and squeeze and lift good the slower you move here the more work you''re going to feel through those quads and glutes as well slow back slow up I know sometimes we feel like we want to get that heart rate going but sometimes we need this slow movement is going to give us even more benefits to support us for those fast movements later' - 'through the pelvic floor and the Deep transverse abdominis woo my legs are burning hopefully yours are okay right away we''re adding a calf raise press it out hold it out there lower the heels lift keep the heels squeezed to come back in up lower lift come back in check your neutral still have that space press lower lift when we''re in this turned out position sometimes we can feel our weight moving into different spots try to keep equal weight kind of towards that second toe good four more full calf raise three two press it out one we''re going to stay here 10 lower and lifts 9 8 7 six it burns five 4 3 2 one kiss those heels back together come halfway in pulse it down nice wide pull squeezing the heels 8 7 6 5 4' - 'one find that lengthened position you''re going to lift up slide those shoulder blades down the back lift up towards the sky big inhale and exhale back and away left hand to Center turn back towards me and lift I apologize if you''re not on the same side when you''re facing me other arm to Center bottom arm all right we''re gonna flip around to do the same thing on the other side so lifting up tall place that hand in front of your shoulder we''re going up and over long spine and then lift shoulders down again up and over and then you use that oblique to lift up exhale lift and lengthen open the spine Flex oblique to come up and three two and one good full mermaid now up and over turn to face the ground separate those arms' pipeline_tag: sentence-similarity library_name: sentence-transformers metrics: - cosine_accuracy@1 - cosine_accuracy@3 - cosine_accuracy@5 - cosine_accuracy@10 - cosine_precision@1 - cosine_precision@3 - cosine_precision@5 - cosine_precision@10 - cosine_recall@1 - cosine_recall@3 - cosine_recall@5 - cosine_recall@10 - cosine_ndcg@10 - cosine_mrr@10 - cosine_map@100 model-index: - name: SentenceTransformer based on Snowflake/snowflake-arctic-embed-m results: - task: type: information-retrieval name: Information Retrieval dataset: name: Unknown type: unknown metrics: - type: cosine_accuracy@1 value: 0.45 name: Cosine Accuracy@1 - type: cosine_accuracy@3 value: 0.69 name: Cosine Accuracy@3 - type: cosine_accuracy@5 value: 0.73 name: Cosine Accuracy@5 - type: cosine_accuracy@10 value: 0.86 name: Cosine Accuracy@10 - type: cosine_precision@1 value: 0.45 name: Cosine Precision@1 - type: cosine_precision@3 value: 0.22999999999999995 name: Cosine Precision@3 - type: cosine_precision@5 value: 0.146 name: Cosine Precision@5 - type: cosine_precision@10 value: 0.08599999999999998 name: Cosine Precision@10 - type: cosine_recall@1 value: 0.45 name: Cosine Recall@1 - type: cosine_recall@3 value: 0.69 name: Cosine Recall@3 - type: cosine_recall@5 value: 0.73 name: Cosine Recall@5 - type: cosine_recall@10 value: 0.86 name: Cosine Recall@10 - type: cosine_ndcg@10 value: 0.6469151250217234 name: Cosine Ndcg@10 - type: cosine_mrr@10 value: 0.5798452380952381 name: Cosine Mrr@10 - type: cosine_map@100 value: 0.5876971664734822 name: Cosine Map@100 --- # SentenceTransformer based on Snowflake/snowflake-arctic-embed-m This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Snowflake/snowflake-arctic-embed-m](https://huggingface.co/Snowflake/snowflake-arctic-embed-m). It maps sentences & paragraphs to a 768-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:** [Snowflake/snowflake-arctic-embed-m](https://huggingface.co/Snowflake/snowflake-arctic-embed-m) - **Maximum Sequence Length:** 512 tokens - **Output Dimensionality:** 768 dimensions - **Similarity Function:** Cosine Similarity ### Model Sources - **Documentation:** [Sentence Transformers Documentation](https://sbert.net) - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, '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: ```bash 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("bsmith3715/legal-ft-demo_final") # Run inference sentences = [ 'What muscle groups are primarily engaged during the lunge exercise described in the context?', "hips and we're gonna lunge it down the\nweight is going to feel a little bit\nlight we're working more stabilizers to\nstart here\nstabilizers through the hips ankles\nknees\nall the things okay lunge it down I like\nto put my foot right up against that\nedge\nto help me have a nice grip toes are off\nof the carriage\nand then coming up to squeeze up on a\nstraight standing leg squeeze up through\nthe glute\ndown\nand\nsqueeze and lift good\nthe slower you move here the more work\nyou're going to feel through those quads\nand glutes as well\nslow back\nslow up\nI know sometimes we feel like we want to\nget that heart rate going\nbut sometimes we need this slow movement\nis going to give us even more benefits\nto support us for those fast movements\nlater", "one\nfind that lengthened position you're\ngoing to lift up slide those shoulder\nblades down the back lift up towards the\nsky big inhale\nand exhale back and away\nleft hand to Center\nturn back towards me and lift I\napologize if you're not on the same side\nwhen you're facing me\nother arm to Center bottom arm all right\nwe're gonna flip around\nto do the same thing on the other side\nso lifting up tall\nplace that hand in front of your\nshoulder we're going up and over long\nspine and then lift shoulders down again\nup and over\nand then you use that oblique to lift up\nexhale lift\nand lengthen\nopen the spine Flex\noblique to come up\nand three\ntwo\nand one\ngood full mermaid now up and over turn\nto face the ground separate those arms", ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 768] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ``` ## Evaluation ### Metrics #### Information Retrieval * Evaluated with [InformationRetrievalEvaluator](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) | Metric | Value | |:--------------------|:-----------| | cosine_accuracy@1 | 0.45 | | cosine_accuracy@3 | 0.69 | | cosine_accuracy@5 | 0.73 | | cosine_accuracy@10 | 0.86 | | cosine_precision@1 | 0.45 | | cosine_precision@3 | 0.23 | | cosine_precision@5 | 0.146 | | cosine_precision@10 | 0.086 | | cosine_recall@1 | 0.45 | | cosine_recall@3 | 0.69 | | cosine_recall@5 | 0.73 | | cosine_recall@10 | 0.86 | | **cosine_ndcg@10** | **0.6469** | | cosine_mrr@10 | 0.5798 | | cosine_map@100 | 0.5877 | ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 200 training samples * Columns: sentence_0 and sentence_1 * Approximate statistics based on the first 200 samples: | | sentence_0 | sentence_1 | |:--------|:-----------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------| | type | string | string | | details | | | * Samples: | sentence_0 | sentence_1 | |:-----------------------------------------------------------------------------------------------|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | What type of spring is the instructor using for the workout? | hi guys thanks for joining me today we
have a really fun really challenging
workout for you today before we get
started don't forget to like share
subscribe feel free to leave me those
super likes I really appreciate you guys
joining me for these workouts we're
going to get started setting up foot
bars all the way down I'm gonna go on to
one red Spring today which is going to
be one heavy spring on my reformer again
I'm gonna go really heavy for my arms
today if this is way too much for you
guys you can do a blue instead of a red
or a medium instead of a heavy spring
again it is going to be very heavy for
arms so feel free to change as needed we
are going to start first by straddling
your reformers your feet are on the | | What should participants do if the red spring is too heavy for them? | hi guys thanks for joining me today we
have a really fun really challenging
workout for you today before we get
started don't forget to like share
subscribe feel free to leave me those
super likes I really appreciate you guys
joining me for these workouts we're
going to get started setting up foot
bars all the way down I'm gonna go on to
one red Spring today which is going to
be one heavy spring on my reformer again
I'm gonna go really heavy for my arms
today if this is way too much for you
guys you can do a blue instead of a red
or a medium instead of a heavy spring
again it is going to be very heavy for
arms so feel free to change as needed we
are going to start first by straddling
your reformers your feet are on the | | What is the initial position described for starting the workout on the reformers? | are going to start first by straddling
your reformers your feet are on the
floor we're going to take our hands to
our shoulder blocks we're just going to
do a quick stretch before we get moving
so I'm going to bend my knees slightly
I'm going to inhale press my Carriage
out let my chest drop down in between my
arms and then my exhale I'm going to
tuck my pelvis around through my spine
to come back in inhale press out let
your chest drop down X tail tuck around
to come in we have two more again after
this we're going to get a really good
workout in today last one
and then round and come in all right now
once we bring it back in we're going to
take our knees onto our carriages and
then our hands are going to go into our | * Loss: [MatryoshkaLoss](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters: ```json { "loss": "MultipleNegativesRankingLoss", "matryoshka_dims": [ 768, 512, 256, 128, 64 ], "matryoshka_weights": [ 1, 1, 1, 1, 1 ], "n_dims_per_step": -1 } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: steps - `per_device_train_batch_size`: 10 - `per_device_eval_batch_size`: 10 - `num_train_epochs`: 10 - `multi_dataset_batch_sampler`: round_robin #### All Hyperparameters
Click to expand - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: steps - `prediction_loss_only`: True - `per_device_train_batch_size`: 10 - `per_device_eval_batch_size`: 10 - `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`: 5e-05 - `weight_decay`: 0.0 - `adam_beta1`: 0.9 - `adam_beta2`: 0.999 - `adam_epsilon`: 1e-08 - `max_grad_norm`: 1 - `num_train_epochs`: 10 - `max_steps`: -1 - `lr_scheduler_type`: linear - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.0 - `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 - `use_ipex`: False - `bf16`: False - `fp16`: False - `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} - `tp_size`: 0 - `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} - `deepspeed`: None - `label_smoothing_factor`: 0.0 - `optim`: adamw_torch - `optim_args`: None - `adafactor`: False - `group_by_length`: False - `length_column_name`: length - `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 - `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`: False - `neftune_noise_alpha`: None - `optim_target_modules`: None - `batch_eval_metrics`: False - `eval_on_start`: False - `use_liger_kernel`: False - `eval_use_gather_object`: False - `average_tokens_across_devices`: False - `prompts`: None - `batch_sampler`: batch_sampler - `multi_dataset_batch_sampler`: round_robin
### Training Logs | Epoch | Step | cosine_ndcg@10 | |:-----:|:----:|:--------------:| | 1.0 | 20 | 0.6130 | | 2.0 | 40 | 0.6454 | | 2.5 | 50 | 0.6445 | | 3.0 | 60 | 0.6498 | | 4.0 | 80 | 0.6507 | | 5.0 | 100 | 0.6463 | | 6.0 | 120 | 0.6433 | | 7.0 | 140 | 0.6461 | | 7.5 | 150 | 0.6409 | | 8.0 | 160 | 0.6417 | | 9.0 | 180 | 0.6425 | | 10.0 | 200 | 0.6469 | ### Framework Versions - Python: 3.13.2 - Sentence Transformers: 4.1.0 - Transformers: 4.51.3 - PyTorch: 2.7.0+cpu - Accelerate: 1.7.0 - Datasets: 3.6.0 - Tokenizers: 0.21.1 ## 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", } ``` #### MatryoshkaLoss ```bibtex @misc{kusupati2024matryoshka, title={Matryoshka Representation Learning}, author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi}, year={2024}, eprint={2205.13147}, archivePrefix={arXiv}, primaryClass={cs.LG} } ``` #### 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} } ```