i486
commented
1 week ago I obtained following using your values. Try removing <|begin_of_text|> as it should be handled internally by llama.
Additionally, f16_kv=True appears to be deprecated since version 0.2.24
RNFL Clock Hours in ONH and RNFL OU Analysis:
In the CIRRUS HD-OCT system, RNFL Clock hours refer to a specific analysis of the retinal nerve fiber layer (RNFL) thickness measurements. This analysis is part of the ONH and RNFL OU Analysis.
The RNFL Clock hours analysis provides information on the thickness of the RNFL at different clock hour positions around the optic disc. The clock hour positions are:
- Clock hour 1: Superior nasal
- Clock hour 2: Superior temporal
- Clock hour 3: Inferior temporal
- Clock hour 4: Inferior nasal
This analysis helps to assess the symmetry and thickness of the RNFL, which can be useful in diagnosing and monitoring various retinal diseases.
Additional Information:
The RNFL Clock hours analysis is based on a reference database of healthy eyes, which provides a range of normal values for each clock hour position. The system compares the patient's measurements to these reference values to provide an assessment of their RNFL thickness and symmetry.
Tag: #CIRRUSHD-OCT
gformcreation
Prerequisites
Please answer the following questions for yourself before submitting an issue.
Expected Behavior
The model should have provided a normal response.
Current Behavior
The model is providing just random garbage values.
Environment and Context
python --version Python 3.11.9nvidia-smi Sat Jul 6 12:18:14 2024 +-----------------------------------------------------------------------------------------+ | NVIDIA-SMI 555.85 Driver Version: 555.85 CUDA Version: 12.5 | |-----------------------------------------+------------------------+----------------------+ | GPU Name Driver-Model | Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. | | | | MIG M. | |=========================================+========================+======================| | 0 Tesla T4 TCC | 00000001:00:00.0 Off | Off | | N/A 48C P0 26W / 70W | 11436MiB / 16384MiB | 0% Default | | | | N/A | +-----------------------------------------+------------------------+----------------------+ +-----------------------------------------------------------------------------------------+ | Processes: | | GPU GI CI PID Type Process name GPU Memory | | ID ID Usage | |=========================================================================================| | 0 N/A N/A 10128 C ...rograms\Python\Python311\python.exe 11416MiB | +-----------------------------------------------------------------------------------------+
pip show llama-cpp-python Name: llama_cpp_python Version: 0.2.81 Summary: Python bindings for the llama.cpp library Home-page: Author: Author-email: Andrei Betlen <abetlen@gmail.com> License: MIT Location: C:\Users\user123\AppData\Local\Programs\Python\Python311\Lib\site-packages Requires: diskcache, jinja2, numpy, typing-extensions Required-by: FlashRankllama_model_loader: loaded meta data with 22 key-value pairs and 291 tensors from C:\Users\user123\Downloads\Meta-Llama-3-8B-Instruct.Q5_K_M.gguf (version GGUF V3 (latest)) llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output. llama_model_loader: - kv 0: general.architecture str = llama llama_model_loader: - kv 1: general.name str = models llama_model_loader: - kv 2: llama.block_count u32 = 32 llama_model_loader: - kv 3: llama.context_length u32 = 8192 llama_model_loader: - kv 4: llama.embedding_length u32 = 4096 llama_model_loader: - kv 5: llama.feed_forward_length u32 = 14336 llama_model_loader: - kv 6: llama.attention.head_count u32 = 32 llama_model_loader: - kv 7: llama.attention.head_count_kv u32 = 8 llama_model_loader: - kv 8: llama.rope.freq_base f32 = 500000.000000 llama_model_loader: - kv 9: llama.attention.layer_norm_rms_epsilon f32 = 0.000010 llama_model_loader: - kv 10: general.file_type u32 = 17 llama_model_loader: - kv 11: llama.vocab_size u32 = 128256 llama_model_loader: - kv 12: llama.rope.dimension_count u32 = 128 llama_model_loader: - kv 13: tokenizer.ggml.model str = gpt2 llama_model_loader: - kv 14: tokenizer.ggml.pre str = llama-bpe llama_model_loader: - kv 15: tokenizer.ggml.tokens arr[str,128256] = ["!", "\"", "#", "$", "%", "&", "'", ... llama_model_loader: - kv 16: tokenizer.ggml.token_type arr[i32,128256] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... llama_model_loader: - kv 17: tokenizer.ggml.merges arr[str,280147] = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "... llama_model_loader: - kv 18: tokenizer.ggml.bos_token_id u32 = 128000 llama_model_loader: - kv 19: tokenizer.ggml.eos_token_id u32 = 128001 llama_model_loader: - kv 20: tokenizer.chat_template str = {% set loop_messages = messages %}{% ... llama_model_loader: - kv 21: general.quantization_version u32 = 2 llama_model_loader: - type f32: 65 tensors llama_model_loader: - type q5_K: 193 tensors llama_model_loader: - type q6_K: 33 tensors llm_load_vocab: special tokens cache size = 256 llm_load_vocab: token to piece cache size = 0.8000 MB llm_load_print_meta: format = GGUF V3 (latest) llm_load_print_meta: arch = llama llm_load_print_meta: vocab type = BPE llm_load_print_meta: n_vocab = 128256 llm_load_print_meta: n_merges = 280147 llm_load_print_meta: n_ctx_train = 8192 llm_load_print_meta: n_embd = 4096 llm_load_print_meta: n_head = 32 llm_load_print_meta: n_head_kv = 8 llm_load_print_meta: n_layer = 32 llm_load_print_meta: n_rot = 128 llm_load_print_meta: n_swa = 0 llm_load_print_meta: n_embd_head_k = 128 llm_load_print_meta: n_embd_head_v = 128 llm_load_print_meta: n_gqa = 4 llm_load_print_meta: n_embd_k_gqa = 1024 llm_load_print_meta: n_embd_v_gqa = 1024 llm_load_print_meta: f_norm_eps = 0.0e+00 llm_load_print_meta: f_norm_rms_eps = 1.0e-05 llm_load_print_meta: f_clamp_kqv = 0.0e+00 llm_load_print_meta: f_max_alibi_bias = 0.0e+00 llm_load_print_meta: f_logit_scale = 0.0e+00 llm_load_print_meta: n_ff = 14336 llm_load_print_meta: n_expert = 0 llm_load_print_meta: n_expert_used = 0 llm_load_print_meta: causal attn = 1 llm_load_print_meta: pooling type = 0 llm_load_print_meta: rope type = 0 llm_load_print_meta: rope scaling = linear llm_load_print_meta: freq_base_train = 500000.0 llm_load_print_meta: freq_scale_train = 1 llm_load_print_meta: n_ctx_orig_yarn = 8192 llm_load_print_meta: rope_finetuned = unknown llm_load_print_meta: ssm_d_conv = 0 llm_load_print_meta: ssm_d_inner = 0 llm_load_print_meta: ssm_d_state = 0 llm_load_print_meta: ssm_dt_rank = 0 llm_load_print_meta: model type = 8B llm_load_print_meta: model ftype = Q5_K - Medium llm_load_print_meta: model params = 8.03 B llm_load_print_meta: model size = 5.33 GiB (5.70 BPW) llm_load_print_meta: general.name = models llm_load_print_meta: BOS token = 128000 '<|begin_of_text|>' llm_load_print_meta: EOS token = 128001 '<|end_of_text|>' llm_load_print_meta: LF token = 128 'Ä' llm_load_print_meta: EOT token = 128009 '<|eot_id|>' llm_load_print_meta: max token length = 256 llm_load_tensors: ggml ctx size = 0.27 MiB llm_load_tensors: offloading 32 repeating layers to GPU llm_load_tensors: offloading non-repeating layers to GPU llm_load_tensors: offloaded 33/33 layers to GPU llm_load_tensors: CPU buffer size = 344.44 MiB llm_load_tensors: CUDA0 buffer size = 5115.49 MiB ......................................................................................... llama_new_context_with_model: n_batch is less than GGML_KQ_MASK_PAD - increasing to 32 llama_new_context_with_model: n_ctx = 2048 llama_new_context_with_model: n_batch = 32 llama_new_context_with_model: n_ubatch = 32 llama_new_context_with_model: flash_attn = 0 llama_new_context_with_model: freq_base = 10000.0 llama_new_context_with_model: freq_scale = 1 llama_kv_cache_init: CUDA0 KV buffer size = 256.00 MiB llama_new_context_with_model: KV self size = 256.00 MiB, K (f16): 128.00 MiB, V (f16): 128.00 MiB llama_new_context_with_model: CUDA_Host output buffer size = 0.49 MiB llama_new_context_with_model: CUDA0 compute buffer size = 16.16 MiB llama_new_context_with_model: CUDA_Host compute buffer size = 0.75 MiB llama_new_context_with_model: graph nodes = 1030 llama_new_context_with_model: graph splits = 2 AVX = 1 | AVX_VNNI = 0 | AVX2 = 1 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 1 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 1 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 0 | Model metadata: {'general.name': 'models', 'general.architecture': 'llama', 'llama.block_count': '32', 'llama.context_length': '8192', 'tokenizer.ggml.eos_token_id': '128001', 'general.file_type': '17', 'llama.attention.head_count_kv': '8', 'llama.embedding_length': '4096', 'llama.feed_forward_length': '14336', 'llama.attention.head_count': '32', 'llama.rope.freq_base': '500000.000000', 'llama.attention.layer_norm_rms_epsilon': '0.000010', 'llama.vocab_size': '128256', 'llama.rope.dimension_count': '128', 'tokenizer.ggml.model': 'gpt2', 'tokenizer.ggml.pre': 'llama-bpe', 'general.quantization_version': '2', 'tokenizer.ggml.bos_token_id': '128000', 'tokenizer.chat_template': "{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{% if add_generation_prompt %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}{% endif %}"} Available chat formats from metadata: chat_template.default Guessed chat format: llama-3
Code :-
from langchain_core.callbacks import CallbackManager, StreamingStdOutCallbackHandler from langchain_community.llms import LlamaCpp callback_manager = CallbackManager([StreamingStdOutCallbackHandler()]) llm = LlamaCpp( model_path="C:\Users\user123\Downloads\Meta-Llama-3-8B-Instruct.Q5_K_M.gguf", callback_manager = callback_manager, n_ctx=2048, max_tokens = 4096, n_gpu_layers=-1, verbose=True, top_p = 0.95, min_p = 0.05, frequency_penalty = 0.0, presence_penalty = 0.0, repeat_penalty = 1.1, top_k = 40, f16_kv=True, stop = ["\\"],
temperature = 0,
seed = 2000,
)
prompt_template = """
<|begin_of_text|><|start_header_id|>system<|end_header_id|>
Use the following pieces of information to answer the user's question.
If you don't know the answer, just say that you don't know, don't try to make up an answer.
Context: {context}
Question: {question}
Answer the question and provide additional helpful information,
based on the pieces of information, if applicable. Be succinct.
End the response with tag which is must.
Responses should be properly formatted and structured so that it can be easily read.
Helpful and the most accurate response from the context and relevant to question :- <|eot_id|>
<|start_header_id|>assistant<|end_header_id|>
"""
prompt = PromptTemplate(
template=prompt_template, input_variables=["context", "question"]
)
[1m> Entering new StuffDocumentsChain chain...[0m
[1m> Entering new LLMChain chain...[0m Prompt after formatting: [32;1m[1;3m <|begin_of_text|><|start_header_id|>system<|end_header_id|>
Use the following pieces of information to answer the user's question. If you don't know the answer, just say that you don't know, don't try to make up an answer.
Context: Instructions for Use Attachment B Reference Database 2 for Healthy Eyes Study–CIRRUS 6000 CIRRUS HD-OCT B.8 ONH and RNFL Images 2660021178037 Rev C 465 / 488B.7.2.1 Ganglion Cell Data Type Variablen Mean (SD) 95% CI Min Max Ganglion Cell Thickness Average Thickness (µm) 826 80.46 (6.36) [80.02, 80.89] 61.00 100.00 Temporal Superior (µm) 826 79.60 (6.25) [79.17, 80.03] 60.00 100.00 Superior (µm) 826 80.98 (6.73) [80.52, 81.44] 59.00 102.00 Nasal Superior (µm) 826 82.16 (7.12) [81.68, 82.65] 58.00 100.00 Nasal Inferior (µm) 826 80.49 (7.13) [80.00, 80.97] 59.00 99.00 Inferior (µm) 826 78.84 (6.83) [78.37, 79.30] 58.00 100.00 Temporal Inferior (µm) 826 80.71 (6.36) [80.28, 81.15] 61.00 103.00 Table 93: Summary of variables, Macular Cube 200x200 B.8ONH and RNFL Images NOTE! These parameters are adjusted for age and optic disc area. Measurement Parameters Analyses RNFL Thickness • RNFL Summary Parameters: –Average RNFL Thickness –Temporal Average RNFL Thickness –Superior Average RNFL Thickness –Nasal Average RNFL Thickness –Inferior Average RNFL Thickness –Inter-eye symmetry (RNFL) –Clock hour parameters –TSNIT Profiles –Neuro-Retinal Rim –RNFL Thickness –RNFL Deviation Map• ONH/RNFL OU • Panomap • Single Eye Summary • Wellness Report ONH Features • Rim Area (mm2) • Average Cup-to-Disc Ratio • Vertical Cup-to-Disc Ratio • Cup Volume (mm3)• ONH/RNFL OU • Panomap • Single Eye Summary Table 94: RNFL thickness and ONH Parameters and Analyses with RDB Data Also see 2Analyze ONH/RNFL OU [ } 276]
Instructions for Use 8 Analyzing Exam Data and Creating Reports CIRRUS HD-OCT 8.2 Posterior Segment Scan Analysis 2660021178037 Rev C 279 / 488# Symbol Explanation Rotation Tool Changes the angle of the ONH spoke. 8.2.3.3.3 Analyzing ONH and RNFL The ONH and RNFL OU Analysis uses two kinds of thickness measurements: •RNFL grid When you move the RNFL grid, the thickness maps, deviation maps, and ONH calculations update automatically. •Super-pixels A total of 50 x 50 (2500) super-pixels are analyzed (optic disc excluded). RNFL Thickness Maps report thickness by showing blue or green for thinner areas and yellow or red for thicker areas (the optic disc appears solid blue). Deviation Maps compare the reference range for healthy eyes for the patient's age and show yellow and red areas that are thinner than 95% and 99% of the (age and optic disc size-adjusted) population, respectively. Interpretation Considerations For some patients, deviation maps can show decrease due to reasons other than pathology. such as: •The patient has strongly myopic or hyperopic eyes •The patient has split-bundle anatomy •The patient has a tilted RNFL bundle pattern If the patient's temporal RNFL that is very thin or absent, the maps might show thickened RNFL. CIRRUS™ HD-OCT compares the RNFL thickness and symmetry of the scan(s) with the reference range for healthy eyes for the patient's age.
Attachment B Reference Database 2 for Healthy Eyes Study–CIRRUS 6000 Instructions for Use CIRRUS HD-OCT B.8 ONH and RNFL Images 466 / 488 2660021178037 Rev CB.8.1 ONH Parameters The data originally collected for the database was analyzed to create the reference ranges for the following ONH parameters: •Rim Area (mm2) •Average Cup to Disc Ratio (square root of cup area over disc area) •Vertical Cup to Disc Ratio (cup height over disc height) •Cup Volume (mm3) B.8.2 RNFL Parameters This study determined the reference range for healthy eyes for the following the retinal nerve fiber layer (RNFL) parameters in healthy subjects ages 18 to 88: •Average RNFL Thickness •Temporal Average RNFL Thickness •Superior Average RNFL Thickness •Nasal Average RNFL Thickness •Inferior Average RNFL Thickness •Inter-eye symmetry (RNFL) •Clock hour parameters •TSNIT Profiles •Neuro-Retinal Rim •RNFL Thickness •RNFL Deviation Map B.8.3 ONH and RNFL Data Type Variablen Mean (SD) 95% CI Min Max ONH Rim Area (mm2) 854 1.30 (0.23) [1.28, 1.31] 0.74 2.21 Average Cup/Disc Ratio 854 0.46 (0.17) [0.45, 0.47] 0.06 0.74 Vertical C/D Ratio 854 0.44 (0.16) [0.43, 0.45] 0.05 0.75 Cup Volume (mm3) 854 0.13 (0.12) [0.12, 0.14] 0.00 0.70 RNFL Thickness Average RNFL Thickness (µm) 854 93.19 (9.28) [92.57, 93.81] 69.00 126.00 Temporal (µm) 854 64.04 (11.53) [63.26, 64.81] 40.00 120.00 Superior (µm) 854 115.05 (15.92) [113.98, 116.12] 72.00 162.00 Nasal (µm) 854 72.35 (11.54) [71.58, 73.13] 41.00 115.00 Inferior (µm) 854 121.37 (16.08) [120.29, 122.45] 72.00 188.00 Clock hour 1 (µm) 854 105.56 (23.11) [104.01, 107.11] 51.00 185.00 Clock hour 2 (µm) 854 88.60 (18.10) [87.39, 89.82] 43.00 160.00 Clock hour 3 (µm) 854 60.73 (10.53) [60.02, 61.44] 38.00 103.00 Clock hour 4 (µm) 854 67.78 (13.79) [66.85, 68.70] 37.00 129.00 Question: What is RNFL Clock hours in ONH and RNFL OU Analysis?
Answer the question and provide additional helpful information, based on the pieces of information, if applicable. Be succinct. End the response with tag which is must.
Responses should be properly formatted and structured so that it can be easily read. Helpful and the most accurate response from the context and relevant to question :- <|eot_id|> <|start_header_id|>assistant<|end_header_id|> [0m Llama.generate: prefix-match hit • 2 }
} B } } } } } }
} F } } } } } } } } } } } [C
}
} } • } } } 2. } } } } } } } } } } } } } } } } } } } } } } } } } } } } } • } } } } }
} } } }
} } } . } . • } & } } . } • • • } } • } } • } } •
llama_print_timings: load time = 124.22 ms llama_print_timings: sample time = 206.91 ms / 224 runs ( 0.92 ms per token, 1082.61 tokens per second) llama_print_timings: prompt eval time = 19358.81 ms / 1776 tokens ( 10.90 ms per token, 91.74 tokens per second) llama_print_timings: eval time = 7789.25 ms / 224 runs ( 34.77 ms per token, 28.76 tokens per second) llama_print_timings: total time = 27901.02 ms / 2000 tokens
[1m> Finished chain.[0m
[1m> Finished chain.[0m {'query': 'What is RNFL Clock hours in ONH and RNFL OU Analysis?', 'result': '•\n2\n\xa0 }\n\n\xa0 }\xa0 B\n }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\n\n\n\n\n }\xa0 F\n\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 [C\n\n\xa0}\n\n\xa0 }\xa0 }\xa0•\n\xa0 }\n\xa0 }\xa0 }\xa02.\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0 }\xa0}\xa0•\n\xa0\xa0\xa0 }\xa0\xa0\xa0\xa0\xa0\xa0\xa0 }\xa0\xa0 }\xa0 }\xa0 }\n\n\xa0 }\xa0 }\xa0 }\xa0}\n\n\xa0 }\xa0 }\xa0\xa0 }\xa0.\xa0 }\xa0\xa0.\n\xa0•\xa0 }\xa0&\xa0 }\xa0 }\xa0.\xa0 }\xa0•\n•\n•\n }\xa0 }\xa0•\n\xa0 }\xa0}\xa0•\n\xa0 }\xa0 }\xa0•\n\n', 'source_documents': [Document(page_content='Instructions for Use Attachment B Reference Database 2 for Healthy Eyes Study–CIRRUS 6000\nCIRRUS HD-OCT B.8 ONH and RNFL Images\n2660021178037 Rev C 465 / 488B.7.2.1 Ganglion Cell Data\nType\nVariablen Mean (SD) 95% CI Min Max\nGanglion Cell Thickness\nAverage Thickness (µm) 826 80.46 (6.36) [80.02, 80.89] 61.00 100.00\nTemporal Superior (µm) 826 79.60 (6.25) [79.17, 80.03] 60.00 100.00\nSuperior (µm) 826 80.98 (6.73) [80.52, 81.44] 59.00 102.00\nNasal Superior (µm) 826 82.16 (7.12) [81.68, 82.65] 58.00 100.00\nNasal Inferior (µm) 826 80.49 (7.13) [80.00, 80.97] 59.00 99.00\nInferior (µm) 826 78.84 (6.83) [78.37, 79.30] 58.00 100.00\nTemporal Inferior (µm) 826 80.71 (6.36) [80.28, 81.15] 61.00 103.00\nTable\xa093: Summary of variables, Macular Cube 200x200\nB.8ONH and RNFL Images\nNOTE!\xa0These parameters are adjusted for age and optic disc\narea.\nMeasurement Parameters Analyses\nRNFL Thickness • RNFL Summary Parameters:\n–Average RNFL Thickness\n–Temporal Average RNFL\nThickness\n–Superior Average RNFL\nThickness\n–Nasal Average RNFL\nThickness\n–Inferior Average RNFL\nThickness\n–Inter-eye symmetry (RNFL)\n–Clock hour parameters\n–TSNIT Profiles\n–Neuro-Retinal Rim\n–RNFL Thickness\n–RNFL Deviation Map• ONH/RNFL OU\n• Panomap\n• Single Eye Summary\n• Wellness Report\nONH Features • Rim Area (mm2)\n• Average Cup-to-Disc Ratio\n• Vertical Cup-to-Disc Ratio\n• Cup Volume (mm3)• ONH/RNFL OU\n• Panomap\n• Single Eye Summary\nTable\xa094: RNFL thickness and ONH Parameters and Analyses with RDB Data\nAlso see\n2Analyze ONH/RNFL OU [ }\xa0276]', metadata={'source': 'C:\Users\user123\Downloads\content\2660021178037_c_cirrus_6000_11.7_ifu_enus.pdf', 'page': 464, '_id': '077141c62da74d258f7a468285030b8f', '_collection_name': 'document_embeddings', 'relevance_score': 0.9270478}), Document(page_content="Instructions for Use 8 Analyzing Exam Data and Creating Reports\nCIRRUS HD-OCT 8.2 Posterior Segment Scan Analysis\n2660021178037 Rev C 279 / 488# Symbol Explanation\nRotation Tool\n Changes the angle of the ONH spoke.\n8.2.3.3.3 Analyzing ONH and RNFL\nThe ONH and RNFL OU Analysis uses two kinds of thickness\nmeasurements:\n•RNFL grid\nWhen you move the RNFL grid, the thickness maps, deviation\nmaps, and ONH calculations update automatically.\n•Super-pixels\nA total of 50 x 50 (2500) super-pixels are analyzed (optic disc\nexcluded).\nRNFL Thickness Maps report thickness by showing blue or green\nfor thinner areas and yellow or red for thicker areas (the optic disc\nappears solid blue).\nDeviation Maps compare the reference range for healthy eyes for\nthe patient's age and show yellow and red areas that are thinner\nthan 95% and 99% of the (age and optic disc size-adjusted)\npopulation, respectively.\nInterpretation Considerations\nFor some patients, deviation maps can show decrease due to\nreasons other than pathology. such as:\n•The patient has strongly myopic or hyperopic eyes\n•The patient has split-bundle anatomy\n•The patient has a tilted RNFL bundle pattern\nIf the patient's temporal RNFL that is very thin or absent, the maps\nmight show thickened RNFL.\nCIRRUS™ HD-OCT compares the RNFL thickness and symmetry of\nthe scan(s) with the reference range for healthy eyes for the\npatient's age.", metadata={'source': 'C:\Users\user123\Downloads\content\2660021178037_c_cirrus_6000_11.7_ifu_enus.pdf', 'page': 278, '_id': '253d8f9d11694dec87f1b6374ee4baa8', '_collection_name': 'document_embeddings', 'relevance_score': 0.82976836}), Document(page_content='Attachment B Reference Database 2 for Healthy Eyes Study–CIRRUS 6000 Instructions for Use\nCIRRUS HD-OCT B.8 ONH and RNFL Images\n466 / 488 2660021178037 Rev CB.8.1 ONH Parameters\nThe data originally collected for the database was analyzed to\ncreate the reference ranges for the following ONH parameters:\n•Rim Area (mm2)\n•Average Cup to Disc Ratio (square root of cup area over disc\narea)\n•Vertical Cup to Disc Ratio (cup height over disc height)\n•Cup Volume (mm3)\nB.8.2 RNFL Parameters\nThis study determined the reference range for healthy eyes for the\nfollowing the retinal nerve fiber layer (RNFL) parameters in healthy\nsubjects ages 18 to 88:\n•Average RNFL Thickness\n•Temporal Average RNFL Thickness\n•Superior Average RNFL Thickness\n•Nasal Average RNFL Thickness\n•Inferior Average RNFL Thickness\n•Inter-eye symmetry (RNFL)\n•Clock hour parameters\n•TSNIT Profiles\n•Neuro-Retinal Rim\n•RNFL Thickness\n•RNFL Deviation Map\nB.8.3 ONH and RNFL Data\nType\nVariablen Mean (SD) 95% CI Min Max\nONH\nRim Area (mm2) 854 1.30 (0.23) [1.28, 1.31] 0.74 2.21\nAverage Cup/Disc Ratio 854 0.46 (0.17) [0.45, 0.47] 0.06 0.74\nVertical C/D Ratio 854 0.44 (0.16) [0.43, 0.45] 0.05 0.75\nCup Volume (mm3) 854 0.13 (0.12) [0.12, 0.14] 0.00 0.70\nRNFL Thickness\nAverage RNFL Thickness (µm) 854 93.19 (9.28) [92.57, 93.81] 69.00 126.00\nTemporal (µm) 854 64.04 (11.53) [63.26, 64.81] 40.00 120.00\nSuperior (µm) 854 115.05 (15.92) [113.98, 116.12] 72.00 162.00\nNasal (µm) 854 72.35 (11.54) [71.58, 73.13] 41.00 115.00\nInferior (µm) 854 121.37 (16.08) [120.29, 122.45] 72.00 188.00\nClock hour 1 (µm) 854 105.56 (23.11) [104.01, 107.11] 51.00 185.00\nClock hour 2 (µm) 854 88.60 (18.10) [87.39, 89.82] 43.00 160.00\nClock hour 3 (µm) 854 60.73 (10.53) [60.02, 61.44] 38.00 103.00\nClock hour 4 (µm) 854 67.78 (13.79) [66.85, 68.70] 37.00 129.00', metadata={'source': 'C:\Users\user123\Downloads\content\xyz.pdf', 'page': 465, '_id': '0c53f682055f48a7a373a0f5543ba7df', '_collection_name': 'document_embeddings', 'relevance_score': 0.75868905})]}
Quantized model :- https://huggingface.co/QuantFactory/Meta-Llama-3-8B-Instruct-GGUF/tree/main https://huggingface.co/QuantFactory/Meta-Llama-3-8B-Instruct-GGUF/blob/main/Meta-Llama-3-8B-Instruct.Q5_K_M.gguf