Instructions to use johnnas12/e5-galaxy-finetuned with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use johnnas12/e5-galaxy-finetuned with sentence-transformers:

from sentence_transformers import SentenceTransformer

model = SentenceTransformer("johnnas12/e5-galaxy-finetuned")

sentences = [
    "query: How can I hicmergeloops?",
    "passage: WindowMasker mkcounts. Construct WindowMasker unit counts table. **What it does**  This tool runs `stage 1 <https://www.ncbi.nlm.nih.gov/IEB/ToolBox/CPP_DOC/lxr/source/src/app/winmasker/>`_ of the WindowMasker analysis to produce a unit counts file for a genome assembly.",
    "passage: GROMACS simulation. for system equilibration or data collection. .. class:: infomark  **What it does**  This tool performs a molecular dynamics simulation with GROMACS.  _____  .. class:: infomark  **Input**         - GRO structure file.        - Topology (TOP) file.  A variety of other options can also be specified:        - MDP parameter file to take advantage of all GROMACS features. Otherwise, choose parameters through the Galaxy interface. See the `manual`_ for more information on the options.        - Accepting and producing checkpoint (CPT) input/output files, which allows sequential MD simulations, e.g. when performing NVT and NPT equilibration followed by a production simulation.        - Position restraint (ITP) file, useful for equilibrating solvent around a protein.        - Choice of ensemble: NVT or NPT.        - Whether to return trajectory (XTC or TRR) and/or structure (GRO or PDB) files.  .. _`manual`: http://manual.gromacs.org/documentation/2018/user-guide/mdp-options.html  _____           .. class:: infomark  **Output**         - Structure and/or trajectory files as specified in the input.",
    "passage: hicMergeLoops. merge detected loops of different resolutions.. Merge detected loops ====================  This script merges the loop locations of different different resolutions.  Loops need to have the following format:  chr start end chr start end  A merge happens if x and y position of a loop overlaps with x and y position of another loop; all loops are considered as an overlap within +/- the bin size of the lowest resolution. I.e. for a loop with coordinates x and y, the overlap to all other loops is searched for (x - lowest resolution) and (y + lowest resolution). If two or more locations should be merged, the one with the lowest resolution is taken as the merged loop.  Example usage:  `$ hicMergeLoops -i gm12878_10kb.bedgraph gm12878_5kb.bedgraph gm12878_25kb.bedgraph -o merged_result.bedgraph -r 25000`  Please recall: We work with binned data i.e. the lowest resolution is therefore the one where we merge the most bases into one bin. In the above example the lowest resultion is 25kb, the highest resolution is 5kb.   For more information about HiCExplorer please consider our documentation on readthedocs.io_  .. _readthedocs.io: http://hicexplorer.readthedocs.io/en/latest/index.html"
]
embeddings = model.encode(sentences)

similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [4, 4]

Notebooks
Google Colab
Kaggle

e5-galaxy-finetuned / config.json

johnnas12

Upload config.json

fea15bd verified 10 months ago

Raw

History Blame Contribute Delete

954 Bytes

	{
	"_name_or_path": "intfloat/e5-base-v2",
	"architectures": [
	"BertModel"
	],
	"attention_probs_dropout_prob": 0.1,
	"classifier_dropout": null,
	"gradient_checkpointing": false,
	"hidden_act": "gelu",
	"hidden_dropout_prob": 0.1,
	"hidden_size": 768,
	"initializer_range": 0.02,
	"intermediate_size": 3072,
	"layer_norm_eps": 1e-12,
	"max_position_embeddings": 512,
	"model_type": "bert",
	"num_attention_heads": 12,
	"num_hidden_layers": 12,
	"pad_token_id": 0,
	"position_embedding_type": "absolute",
	"torch_dtype": "float32",
	"transformers_version": "4.49.0",
	"type_vocab_size": 2,
	"use_cache": true,
	"vocab_size": 30522,
	"word_embedding_dimension": 768,
	"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
	}